Skip to main content
main-content

Über dieses Buch

This two volume set LNCS 10602 and LNCS 10603 constitutes the thoroughly refereed post-conference proceedings of the Third International Conference on Cloud Computing and Security, ICCCS 2017, held in Nanjing, China, in June 2017.

The 116 full papers and 11 short papers of these volumes were carefully reviewed and selected from 391 submissions. The papers are organized in topical sections such as: information hiding; cloud computing; IOT applications; information security; multimedia applications; optimization and classification.

Inhaltsverzeichnis

Frontmatter

Information Security

Frontmatter

Practical Privacy-Preserving Outsourcing of Large-Scale Matrix Determinant Computation in the Cloud

Large-Scale matrix determinant computation (LMDC) is a common scientific and engineering computational task and has a number of applications. But such computation involves enormous computing resources, which is burdensome for the clients. Cloud computing enables computational resource-constrained clients to economically outsource such computations to the cloud server. In this paper, we investigate the privacy-preserving large-scale matrix determinant computation outsourcing problem, where the clients can outsource LMDC to the untrusted cloud server, relieving the clients from computation burden. We propose a new privacy-preserving algorithm for outsourcing LMDC, which substantially reduces the computation burden on the client side. Our algorithm builds on a series of carefully-designed pseudorandom matrices, which can hide the original matrix from the cloud server with low computational complexity. The extensive security analysis shows that our algorithm is practically-secure, and offers a higher level of privacy protection than the state-of-the-art on LMDC outsourcing. We provide extensive theoretical analysis and experimental evaluation to show its high-efficiency and security compared to the previous works.

Shaojing Fu, Yunpeng Yu, Ming Xu

Probability-p Order-Preserving Encryption

Order-Preserving Encryption (OPE) is an encryption preserving the order relationship of the plaintexts to support efficient range query on ciphertexts. Other than traditional symmetric encryption aiming at absolute security, OPE sacrifices some security for the ability to search on ciphertext. In this paper, we propose a new cryptographic primitive, Probability-p Order-Preserving Encryption (p-OPE), which preserves the order of plaintexts with probability p. When $$p=1$$p=1, p-OPE becomes OPE, thus p-OPE is an extension of OPE. We define and analyse the security and precision of the novel primitive, then we propose a construction of p-OPE and conduct experiments to show its performance. As shown in the theoretical analysis and experiment results, p-OPE can improve the security at the cost of some precision sacrifice.

Ce Yang, Weiming Zhang, Jiachen Ding, Nenghai Yu

DLPDS: Learning Users’ Information Sharing Behaviors for Privacy Default Setting in Recommender System

The proliferation of Internet of things has allowed users to provide preference feedback and maintain profiles in multiple websites, which could indicate tastes covering several kinds of domains but focus on small number of topics. Leveraging all the user information available in several sites or domains may be beneficial for knowing the users better and generating higher-quality recommendations. However, aggregating all users’ information globally could trigger users’ awareness on privacy concerns, which further cause users refuse to share the information so that the recommendation quality is reduced. We provide evidence that a recommender system could mitigate the privacy-integration problem, when it is applied with our novel model, called disclosure-learning privacy default setting (DLPDS), which transfer the pattern of users’ past information sharing behaviors into privacy default settings. The result of the experiment support that our DLPDS model could gain users’ trust and aggregate more users information, and that adapting the privacy default settings to the user information sharing pattern may results in positive feedback that promoting better prediction accuracy of the recommender system.

Hongchen Wu, Huaxiang Zhang

New Construction of Low-Hit-Zone Frequency Hopping Sequence Sets with Optimal Partial Hamming Correlation

A new construction of low-hit-zone (LHZ) frequency hopping sequence (FHS) sets with optimal partial Hamming correlation (PPHC) is proposed in this paper. Our construction yields LHZ FHS sets with optimal PPHC by interleaving techniques, in which short FHSs with good Hamming correlation are used as base sequences while certain appropriate sequences are chosen as shift sequences. The LHZ FHS sets proposed in this paper have new parameters not covered in the literature.

Zhengqian Li, Pinhui Ke, Zhifan Ye

Detecting and Preventing DDoS Attacks in SDN-Based Data Center Networks

Distributed denial-of-service (DDoS) attacks are deemed a serious threat to Internet services. A common solution to mitigate the attacks is to redirect traffic to scrubbing centers (SCs) for traffic classification and DDoS filtering. However, the capacity and locations of SCs should be pre-determined, and traffic redirection to SCs also give rise to extra network footprint and long latency. In this work, we present a solution based on network function virtualization (NFV) to launch scrubbing functions on demand and software-defined networking (SDN) to redirect traffic to these functions. We propose a lightweight probing strategy to identify anomalous traffic and the victim, and allocate virtual scrubbing functions close to the victim to minimize network footprint and network latency. We simulate a proof-of-concept design in Mininet to demonstrate its operation. The evaluation shows 96.6% of DDoS packets can be mitigated with the response time of one second.

Po-Ching Lin, Yu-Ting Hsu, Ren-Hung Hwang

Bayesian Game Based Pseudo Honeypot Model in Social Networks

In this paper, we study applying honeypots to protect social networks against DDoS attacks. Different from previous works that study honeypots for DDoS attacks, we consider attackers are rational and know to optimize attacking strategies based on the defender’s strategy. To deal with such strategic attackers, we propose a novel pseudo honeypot game model following the Bayesian game setting. In addition, we rigorously prove the existence of Bayesian Nash equilibriums (BNEs) and show how to find them in all different cases. Simulations show the BNEs achieved in the games not only reduce energy consumption but also improve efficiency of the defense.

Miao Du, Yongzhong Li, Qing Lu, Kun Wang

Conducting Correlated Laplace Mechanism for Differential Privacy

Recently, differential privacy achieves good trade-offs between data publishing and sensitive information hiding. But in data publishing for correlated data, the independent Laplace noise implemented in current differential privacy preserving methods can be detected and sanitized, reducing privacy level. In prior work, we have proposed a correlated Laplace mechanism (CLM) to remedy this problem. But the concrete steps and detailed parameters to imply CLM and the complete proof has not been discussed. In this paper, we provide the complete proof and specific steps to conduct CLM. Also, we have verified the error of our implement method. Experimental results show that our method can retain small error to generate correlated Laplace noise for large quantities of queries.

Hao Wang, Zhengquan Xu, Lizhi Xiong, Tao Wang

A Symmetric Authenticated Proxy Re-encryption Scheme with Provable Security

In crypto 2013, Dan et al. proposed a symmetric proxy re-encryption scheme based on key homomorphic PRF. It can be used to ensure the data privacy in cloud storage systems. However, it only focuses on preventing a honest-but-curious proxy from learning anything about the encrypted data. Although it can be made to provide integrity without disrupting the key homomorphism property by using MAC then encrypt with counter-mode, it’s not a symmetric authenticated proxy re-encryption scheme because only the data owner can verify the integrity of some encrypted data. In this paper, we propose a symmetric authenticated proxy re-encryption scheme which can prevent a malicious proxy from tampering users’ data. It can update the authentication tag as well as the ciphertext so that any intended user can verify the integrity of the encrypted data.

Zhiniang Peng, Shaohua Tang, Linzhi Jiang

Quantum Secret Sharing in Noisy Environment

As an unavoidable factor of real-world implementation of quantum cryptograph, quantum noise severally affects the security and reliability of the quantum system. In this paper, we study how QSS, an important branch of quantum cryptograph, is affected by noise or decoherence. QSS protocols for sharing classical information and quantum states are studied in four types of noise that usually encountered in real-world, i.e., the bit-flip, phase-flip (phase-damping), depolarizing and amplitude-damping noise, respectively. Two methods are introduced to evaluate the effect of noise. For the QSS protocol sharing classical information, the efficiency for generating secret key is used. Our results show that the efficiencies are quiet different from each other in four types of noise. While for the protocol sharing quantum states, the output states and the state-independent average fidelity are studied, respectively. It indicates that the players will get two different output states in the amplitude-damping noise, but get one output state in the other three types of noise. Besides, the state-independent average fidelity behaves differently from each other. Our study will be helpful for analyzing and improving quantum secure communications protocols in real-world.

Ming-Ming Wang, Zhi-Guo Qu, Mohamed Elhoseny

SimHash-Based Similar Neighbor Finding for Scalable and Privacy-Preserving Service Recommendation

With the ever-increasing number of web services on the Web, various service recommendation techniques, e.g., User-based Collaborative Filtering (UCF) have been developed to alleviate the service selection burden of a target user. In traditional UCF-based recommendation approaches, the first key step is to find the candidate users who have the largest invoked-service intersection with the target user, as a larger invoked-service intersection often means higher probability that two users are similar friends (i.e., neighbors). However, the above similar-neighbor finding process is often time-consuming and vulnerable to privacy, especially when the number of candidate users is huge. In view of these challenges, a scalable and privacy-preserving similar-neighbor finding approach based on SimHash, i.e., SNF SimHash is proposed in this paper. Finally, a set of experiments are conducted on WS-DREAM dataset to validate the feasibility of SNF SimHash . Experiment results show that our proposal can achieve a good tradeoff between accuracy and efficiency while guaranteeing privacy-preservation.

Yanwei Xu, Lianyong Qi

Universally Composable Three-Party Password Authenticated Key Exchange

Three-party password authenticated key exchange (3PAKE) allows two clients, each sharing a password with a trusted server, to establish a session key with the help of the server. It is a quite practical mechanism for establishing secure channels in large communication network. However, most current 3PAKE protocols are analyzed in security models that don’t adequately address protocol composition problem. In this paper, a direct definition of security for 3PAKE within the universal composability framework is proposed, which captures the basic security requirements of the problem and is proven to be stronger than the commonly used security notions. To further justify our formulation of 3PAKE, we prove that a slight variant of a generic 3PAKE protocol by Wang and Hu securely realizes the new security definition.

Qihui Zhang, Xuexian Hu, Jianghong Wei, Wenfen Liu

A High-Capacity Quantum Secret Sharing Protocol Based on Single D-Level Particles

A new quantum secret sharing protocol is proposed to share a private key based on single d-level particles. A generalized definition of capacity is also given to weigh the total efficiency of such QSS protocols. It is shown that the capacity of this protocol is $$ \log_{2} d $$log2d, higher than the ones using single two-level particles (the maximum capacity is 1) and the similar ones proposed by Tavakoli et al. and Karimipour et al. (the capacities are $$ \log_{2} d/d $$log2d/d and $$ \log_{2} d/2 $$log2d/2 respectively). Besides, it is secure against several common attacks and feasible with present-day technology.

Xiang Lin, Juan Xu, Hu Liu, Xuwei Tang, Maomao Fu

An Inconsistency Detection Method for Security Policy and Firewall Policy Based on CSP Solver

Packet filtering in firewall either accepts or denies network packets based upon a set of pre-defined rules called firewall policy. Firewall policy always designed under the instruction of security policy, which is a generic document that outlines the needs for network access permissions. The design of firewall policy should be consistent with security policy.If firewall policy is not consistent with security policy, firewall policy may violate the intentions of security policy, which is the reason that result in critical security vulnerabilities. This paper extends our previous method, which represented security policy and firewall policy as Constraint Satisfaction Problem (CSP) and used a CSP solver Sugar only to verify whether they are consistent. In this paper, we propose a method to detect and resolve inconsistencies of firewall policy and security policy. We have implemented a prototype system to verify our proposed method, experimental results show the effectiveness.

Yi Yin, Yuichiro Tateiwa, Yun Wang, Yoshiaki Katayama, Naohisa Takahashi

Reversible Data Hiding in Encrypted Image Based on Block Classification Scrambling

To improve the security and the quality of decrypted image, this work proposes a reversible data hiding in encrypted image based on block classification scrambling. In image encryption phase, all 8 × 8 blocks are firstly classified into the smooth blocks and texture ones and the corresponding block-type matrix is generated according to their most significant bits (MSB). And then the XOR-encrypted smooth and texture blocks are scrambled by the encryption key, respectively. At last, the encrypted image is generated by scanning the scrambled smooth and texture blocks in order. The risk of content disclosure of encrypted image obtained by the proposed method is reduced since the value and location of each pixel are both protected. Experimental results demonstrate that the probability for lossless recovery of proposed method is 100% even if for texture image Baboon at a high embedding rate of 0.15 bpp.

Fan Chen, Hongjie He, Tinghuai Ma, Shu Yan

A Mutation Approach of Detecting SQL Injection Vulnerabilities

As Internet is increasingly prosperous, Web services become more common in our social life. As users can access pages on the Web directly, Web application plays a vital role in various domains such as e-finance and public-services. Inevitably, it will be followed by unprecedented amount of attacks and exploitations. Amongst all of those attacks, SQL injection attacks have consistently high rank in last years due to corresponding vulnerabilities. It is crucial to checking this vulnerabilities before web services being public. In our paper we present an effective approach for testing, MOSA, and mutation operators set to its underpinning. Using this approach we can produce test inputs that cause executable and malignant SQL statement efficiently. Besides that, we do numerous experiments and the results demonstrate that the mutation approach can detect SQL injection vulnerabilities and generate inputs that bypass web application firewalls.

Yanyu Huang, Chuan Fu, Xuan Chen, Hao Guo, Xiaoyu He, Jin Li, Zheli Liu

A Chaotic Map-Based Authentication and Key Agreement Scheme with User Anonymity for Cloud Computing

Cloud computing is a hot issue mentioned more and more nowadays. Vast information for every domain are stored in cloud servers. Security issue is accompanied with the fast development of cloud application. On the other hand, many authentication schemes with different methods have been presented over the last decades. To guarantee the valid access to remote server, smart cards are used on the client side. Moreover, user anonymity becomes a hot issue in such schemes. We present a chaotic-map based mutual authentication scheme for cloud computing. After our concrete analysis, the new scheme not only makes user anonymous but also performs well in basic aspects compared with recent schemes. The new scheme is practical and efficient via our comparison and it overcomes various attacks and meets security requirements in public opinion. ...

Fan Wu, Lili Xu

An Anonymous User Authentication and Key Distribution Protocol for Heterogenous Wireless Sensor Network

With the development of the Internet of Things (IOT), peoples’ lifestyle and social inter-communication have been changed greatly. As a part of the IOT, wireless sensor network (WSN) also attracts many researchers to pay close attention. In this paper, we investigate one latest scheme (Farash et al.’s scheme) that provides an efficient user authentication and key agreement scheme for heterogeneous wireless sensor network tailored for the Internet of Things environment. However, we find some shortcomings and weaknesses in this scheme. It cannot supply anonymity for users and sensors and cannot protect stolen smart card attack. So our work is to eliminate the threat of cryptographic attack and to improve security.

Xin Zhang, Fengtong Wen

Privacy-Preserving Multi-party Aggregate Signcryption for Heterogeneous Systems

To achieve heterogeneous communication from certificateless cryptography (CLC) to identity-based cryptography (IBC), we present a heterogeneous scheme that enables m senders in the CLC to transmit m message to n receivers in the IBC. In the proposed signcryption scheme, each sender is mapped to a distinct pseudo identity, so the sender’s identity privacy preservation can be guaranteed. At the same time, to ensure the receiver’s identity privacy, the identity information of all authorized recipients is mixed by the Lagrange interpolation polynomial during the signcryption process, which prevents the receiver’s identity from being exposed. Compared with existing schemes, the proposed scheme presents efficient computational overhead and is suitable for heterogeneous environments. In addition, our scheme has the indistinguishability against adaptive chosen ciphertext attacks and existential unforgeability against adaptive chosen-message attacks in the random oracle model.

Shufen Niu, Zhenbin Li, Caifen Wang

An Enhanced Method of Trajectory Privacy Preservation Through Trajectory Reconstruction

Trajectory data of mobile users contain plenty of sensitive spatial and temporal information, and can support many applications through data analysing and mining. However, re-identification attack and inference attack on such data may cause serious personal privacy leakage. Existing privacy preserving techniques cannot protect trajectory privacy well or largely scarify data utility. In view of these issues, in this paper we propose an enhanced trajectory privacy preserving method which can protect the trajectory privacy preferably while maintaining a high utility of the trajectory in data publishing. A mechanism is proposed to protect the privacy through replacing stop points in the trajectory and an effective trajectory reconstruction algorithm is introduced to avoid the mutations of trajectory, and also deal with the possible presence of obstacles around trajectories. The performance of our proposal is comprehensively evaluated on a real trajectory dataset. The results show that our method achieves a high privacy level and improves the utility of trajectory data greatly, compared with the state-of-the-art method.

Yan Dai, Jie Shao

Efficient and Short Identity-Based Deniable Authenticated Encryption

Deniable authentication is an important security requirement for many applications that require user privacy protection, since the sender can deny that he/she has signed the message. Considering the importance of communication efficiency, in this paper, we explore the novel deniable authenticated encryption, which outperforms the existing ones in terms of communication costs and ciphertext size. Our protocol meets all the security requirement of message confidentiality and deniable message authentication. Our protocol is based on identity cryptography and can avoid the public key certificates based public key infrastructure (PKI). Our protocol is provably secure in the random oracle model.

Chunhua Jin, Jianyang Zhao

Towards Fully Homomorphic Encryption From Gentry-Peikert-Vaikuntanathan Scheme

Despite the convenience brought by cloud computing, internet users, meanwhile, are faced with risks of data theft, tampering, forgery, etc. Fully homomorphic encryption (FHE) has the ability to deal with the ciphertext directly, which can solve the problem of data security in cloud computing. Therefore, fully homomorphic encryption (FHE) has been widely used in cloud computing as well as multiparty computing, functional encryption and private information retrieval, etc. However, previous FHE schemes are based on standard (ring) learning with errors (LWE) assumption and the most typical schemes were created by Brakerski (CRYPTO2012) and Gentry-Sahai-Waters (GSW) (CRYPTO2013). Moreover, inspired by the work of Li et al. at ICPADS2016, they made use of Brakerski’s scale-invariant technology and constructed a new FHE scheme with errorless key switching under Dual-First-is-errorless LWE (Dual-Ferr.LWE) problem. Hence, armed with Li et al.’s work, in this paper, we use Gentry-Peikert-Vaikuntanathan’s scheme (i.e., under dual LWE assumption) as building block to construct a FHE scheme. Lastly, under the assumption of decisional learning with errors (LWE), we prove that our scheme is CPA (chosen-plaintext-attack) secure.

Gang Du, Chunguang Ma, Zengpeng Li, Ding Wang

Privacy-Preserving Medical Information Systems Using Multi-authority Content-Based Encryption in Cloud

In the Medical Information Systems (MIS), the patient outsources his e-health records, a dramatically huge amount of health data, to a third party like cloud service provider. The Internet providing host-to-host communication using TCP/IP network topology has not satisfied the growing demands of data processing in MIS. Based on the content-to-consumer paradigm, content-centric networking architecture was proposed for simple easy-to-manage caching features to users. In this paper, we proposed a privacy-preserving e-health records scheme that protects name and content simultaneously. Our proposal has multi-authority without a trusted single or central authority to distribute secret keys, which avoids the key escrow problem and meets the distributed features of MIS. As we know, this scheme is the first multi-authority content-based encryption (MA-CBE). Furthermore, this MA-CBE resists up to (N-1) corrupted authorities collusion attack, and the security is proven to be semantically secure based on the standard decisional bilinear Diffie-Hellman assumption. Our comparison analysis reports that the proposal gives a better performance than other related schemes.

Rui Guo, Xiong Li, Dong Zheng

Detect Storage Vulnerability of User-Input Privacy in Android Applications with Static and Dynamic Analysis

In recent years Android has become the most popular operating system in mobile phone, and a variety of apps bring people great convenience in our daily life and work. Due to the resource constraints in mobile phone and user experience considerations, a large number of private data are stored in the phone itself. Privacy Leaks will bring huge losses to us. EditText, which is designed for Android developers to input the sensitive data (e.g. username, password, search keywords etc.) to the apps, carries much User-Input Privacy (UIP) data. So, whether these UIP data is stored in the phone safely becomes the key to protect the privacy. In this paper, we do the research about the UIP data in EditText widget, and detect whether the data entered by the user is safely stored through static taint analysis and dynamic Smali Instrumentation. Experiments show that some of the apps store the UIP data in EditText at an unsafe location or store them in a weak way, which will bring the risk of privacy leakage.

Li Jiang, Yi Zhuang

Certificateless Cryptography with KGC Trust Level 3 Revisited

This paper revisits the issue of obtaining KGC (Key Generator Center) trust level 3 in certificateless cryptography. The AP (Al-Riyami-Paterson) binding technique can modularly construct the certificateless encryption/signature scheme with trust level 3 from that with trust level 2. However, its security proof has been an open problem. Yang and Tan improved the AP framework by adding extra cryptographic tools: random oracles for security proof in the random oracle model or trapdoor hash functions for security proof in the standard model. This paper aims to prove secure the original AP binding technique. The basic technique for achieving this security proof depends on the improved security model for certificateless encryption (or signature) schemes. As an application example, one key dependent certificateless encryption scheme with both authority trust level 3 and provable security in the standard model is modularly constructed by applying the AP binding framework to one conventional certificateless encryption scheme.

Fei Li, Wei Gao, Dongqing Xie, Chunming Tang

Server-Less Lightweight Authentication Protocol for RFID System

The design of secure authentication protocols for RFID system is still a great challenging problem. Many authentication protocols for RFID have been presented, but most have security flaws. We analyzes the security of scheme proposed by Deng et al., and point out that this scheme can’t resist location tracking attack, and the low efficiency of the reader searches a target tag. Based on this, an improved protocol to overcome the security vulnerability of Deng’s protocol is presented. The formal proof of correctness of the improved protocol is given based on GNY logic which is one of the model logics, and finally experiments shows the improved protocol has the good efficiency of time complexity.

Jing Li, Zhiping Zhou, Ping Wang

Efficient CP-ABE with Non-monotonic Access Structures

Ciphertext policy attribute based encryption (CP-ABE) systems are suitable for supporting access control with complex attribute-based policies. But there is a few CP-ABE systems that support non-monotonic access structures. In this paper we present an efficient CP-ABE construction which supports $$\mathbf {NOT}$$NOT operation as well as m-of-n threshold gates. This construction is proved to be selectively secure under decisional bilinear Diffie-Hellman assumption. The comparison between our scheme and other similar systems shows that our construction is an efficient and expressive CP-ABE construction that supports the negation of attributes.

Yong Cheng, Huaizhe Zhou, Jun Ma, Zhiying Wang

Image Authentication Based on Least Significant Bit Hiding and Double Random Phase Encoding Technique

In this paper, we proposed an image authentication method based on least significant bit (LSB) hiding and double random phase encoding (DRPE) technique. An image that needs to be authenticated is first encrypted using DRPE algorithm. All amplitude information in this encrypted image is discarded and the phase mask is reserved. Then, the phase image is converted into a binary image by setting phase values less than zero as zero and other values as one. The binary image is hided in a host image with the LSB hiding approach. The recipient extracts the hiding binary image from the host image and then sets the zero-values as $$ { - \pi } $$-π and one-values as $$ \uppi $$π. Finally, the retrieved phase mask is decrypted and authenticated using DRPE and non-linear cross correlation algorithm, respectively. The proposed authentication scheme can provide another layer of security because the retrieved phase mask cannot reveal any original image information. Furthermore, the decrypted image from phase mask cannot be visually recognized with naked eyes which can distract the attention of attackers. Experimental results have shown the feasibility of the proposed authentication algorithm.

Faliu Yi, Yousun Jeoung, Ruili Geng, Inkyu Moon

On-Line Intrusion Detection Model Based on Approximate Linear Dependent Condition with Linear Latent Feature Extraction

Most of the intrusion detection models (IDM) are constructed with off-line training data. Time-variance characteristic of the practical network system cannot be embodied in the off-line constructed IDM. On-line updating of the off-line IDM with the valued new samples is very necessary. In this paper, a new on-line instruction detection model based on approximate linear dependent (ALD) condition with linear latent feature extraction is proposed to address this problem. Specifically, the valued samples which can represent drift of the practical network are indentified with ALD and prior knowledge. Then, these selected samples are used to update the off-line IDM based on on-line latent feature extraction method and fast machine learning algorithm with sample-based updating strategy. Experiments based on KDD99 data are used to validate the proposed approach.

Jian Tang, Meijuan Jia, Jian Zhang, Meiying Jia

Differential Direction Adaptive Based Reversible Information Hiding

In order to reversibly hide information in images with high capacity, an self-adaptive method is proposed in this paper by optimally selecting the differential direction. By dividing image into blocks, differences for adjacent pixel pairs in each blocks are computed in multi-directions. Based on the statistics of these differences, the optimal and self-adaptive strategy for increasing embedding capacity is introduced. The experimental results demonstrate that a significant improvement of embedding capacity is achieved, while the qualities of images is maintained.

Feipeng Lin, Bo Wang, Yabin Li

Detection of Jitterbug Covert Channel Based on Partial Entropy Test

Jitterbug is a typical delay-based covert timing channel and supplies reliable covert communication in a passive manner. The existing entropy-based detection scheme based on training samples may suffer from model mismatching, which results in detection performance deterioration. In this paper, a new detection method for Jitterbug based on partial entropy test is proposed. A fixed binning strategy without training samples is used to obtain bins distribution feature. The first-order entropy is calculated for several sets of partial successive bins and the weighted mean is used to calculate the final entropy value to distinguish Jitterbug from legitimate traffic. Furthermore, the influence of detection performance caused by network jitter is also discussed. Experimental results show that the proposed detection method achieves high detection performance and is less affected by network jitter.

Hao Wang, Guangjie Liu, Weiwei Liu, Jiangtao Zhai, Yuewei Dai

Multimedia Applications

Frontmatter

Frame-Deletion Detection for Static-Background Video Based on Multi-scale Mutual Information

Due to enormous free video editing software on the Internet, tampering of digital videos has become very easy. Authenticating the integrity of videos and detecting any video forgery is a big challenge to researchers. In this paper, an algorithm based on the normalized mutual information feature is proposed to detect the frame-deleting videos which are hardly identified by human visual. The proposed method is composed of two parts: feature extraction and abnormal point detection. Firstly, based on information theory, the normalized mutual information is defined on the single scale visual content of adjacent frames. After using the Gaussian pyramid transform on every frame, the description operator of multi-scale normalized mutual information is computed by linear combination. In the stage of discontinuity point detection, video forgery is identified and the tampering point is localized by performing modified generalized ESD test.

Yanjia Zhao, Tianming Pang, Xiaoyun Liang, Zhaohong Li

Learning Based Fast H.264/AVC to HEVC INTRA Video Transcoding for Cloud Media Computing

Cloud video transcoding enable to convert the video standards and properties from one to another so as to adapt to different user end devices and network capacity, especially in sharing massive video contents in cloud environment. High Efficiency Video Coding (HEVC) and H.264/Advanced Video Coding are two recent high performance video coding standards that are widely used and co-existing in video industry. Video transcoding is desirable to bridge the standard gap. To effectively transcode video stream from H.264/AVC to HEVC for higher compression efficiency and meanwhile maintaining low computational complexity, a learning based fast H.264/AVC to HEVC transcoder is proposed for cloud media computing. We firstly analyze the correlation of block partition sizes between these two standards and then present a fast Coding Unit (CU) decision algorithm, in which three levels of binary classifiers are used to predict different CU sizes in HEVC intra coding and the optimal parameters are determined by statistical experiments. The experimental results show that the proposed transcoder achieves 44.3% time saving on average with only negligible quality degradation when compared with the original cascaded transcoder and is also superior than the state-of-the-art benchmarks in terms of complexity reduction and rate-distortion performance.

Yun Zhang, Na Li, Zongju Peng

A Perceptual Encryption Scheme for HEVC Video with Lossless Compression

Aiming to protect the video content and facilitate online video consumption, a perceptual encryption scheme is proposed for high efficiency video coding (HEVC) video. Based on RC4 algorithm, a key stream generation method is constructed, whose proportion of “1” and “0” can be regulated. During HEVC encoding, four kinds of syntax elements including motion vector difference (MVD)’ sign, MVD’s amplitude, sign of the luma residual coefficient and sign of the chroma residual coefficient, are encrypted by the regulated key stream. Experimental results and analysis show that the proposed scheme has good perceptual protection for the video content, and some advantages such as low computational cost, format-compliance and no bitrate increase can be achieved. It provides an effective resolution for the paid video-on-demand services in smart cities.

Juan Chen, Fei Peng, Min Long

Spatial-Temporal Correlation Based Multi-resolution Hybrid Motion Estimation for Frame Rate Up-Conversion

This paper presents a method applied in frame rate up-conversion (FRUC) to estimate the motion vector (MV) based on spatial-temporal motion vectors, which have a high correlation with the interpolated block. Meanwhile, a hierarchical structure is employed to generate a series of low-resolution images for motion estimation (ME) aimed to reduce computation complexity. After using ME in the lowest resolution image, there are several relatively accurate MVs picked to transmit down to the next level for further calculation. Then the initial motion vector field (MVF) is generated by calculating the sums of absolute difference (SAD) of the candidate MVs. To refine the MVF, an outliers detecting criterion is set up to smooth the MVF and obtain a more accurate MVF at the same time. The motion compensated (MC) will not perform until the termination of MVF refinement level-by-level. Experimental reveals that the proposed method find a tradeoff between accuracy and complexity, because the peak signal-to-noise ratio (PSNR) is promoted to 4.69 dB and the processing time is quite fast.

Bingyu Ji, Ran Li, Changan Wu

A Robust Seam Carving Forgery Detection Approach by Three-Element Joint Density of Difference Matrix

Seam carving is a popular content-aware image retargeting technique. However, it can also be used for malicious purposes such as object removal. In this paper, a robust blind forensics approach is proposed for seam-carved forgery detection. Since insignificant pixels along seams are removed for image resizing, the spatial neighborhood relations among pixels will be significantly changed, especially in smooth regions. Thus, joint density is exploited to model the change of spatially adjacent pixels’ distribution caused by seam carving, even in the case of low scaling ratios. Specifically, three-element joint density of difference matrix is computed to form general forensics features (GTJD). The GTJD features are combined with existing energy and noise features exacted in LBP domain for classification. Experimental results show that the proposed approach achieves better accuracies for both uncompressed images and JPEG images with different scaling ratios.

Wenwu Gu, Gaobo Yang, Dengyong Zhang, Ming Xia

Chinese Remainder Theorem-Based Secret Image Sharing for (k, n) Threshold

In comparison with Shamir’s original polynomial-based secret image sharing (SIS), Chinese remainder theorem-based SIS (CRTSIS) overall has the advantages of lossless recovery, low recovery computation complexity and no auxiliary encryption. Traditional CRTSIS methods generally suffer from no (k, n) threshold, lossy recovery, ignoring the image characteristics and auxiliary encryption. Based on the analysis of image characteristics and SIS, in this paper we propose a CRTSIS method for (k, n) threshold, through dividing the gray image pixel values into two intervals corresponding to two available mapping intervals. Our method realizes (k, n) threshold and lossless recovery for gray image without auxiliary encryption. Analysis and experiments are provided to indicate the effectiveness of the proposed method.

Xuehu Yan, Yuliang Lu, Lintao Liu, Song Wan, Wanmeng Ding, Hanlin Liu

Kernel Searching Strategy for Recommender Searching Mechanism

A trust-aware recommender system (TARS) is widely used in social media to find useful information. Recommender searching mechanism is an important research issue in TARS. We propose a new searching strategy for recommender searching mechanism of TARS, which named kernel searching strategy. A kernel, which consists of hub nodes of the trust network, is involved in trust propagations. The kernel can be obtained from node degree or node betweenness, take these hub nodes as active users and then finds the recommenders via trust propagations from the kernel, most of the nodes in the network will be covered. Comparing the results of these two methods, the coverage rate of these hub nodes which is obtained from the node degree is almost less than that obtained from the node betweenness. To get better coverage rate, we take both degree and betweenness into consideration. The results show that the combination can get better coverage rate only compared with the node degree. However, the combination has better convergence effect compared with the node betweenness.

Li Zhou, Weiwei Yuan, Kangya He, Chenliang Li, Qiang Li

A Robust Quantum Watermark Algorithm Based on Quantum Log-Polar Images

Copyright protection for quantum image is an important research branch of quantum information technology. In this paper, based on quantum log-polar image (QUALPI), a new quantum watermark algorithm is proposed to better protect copyright of quantum image. In order to realize quantum watermark embedding, the least significant qubit (LSQb) of quantum carrier image is replaced by quantum watermark image. Compared to previous quantum watermark algorithms, the new algorithm effectively utilizes two important properties of log-polar sampling, i.e., rotation and scale invariances. These invariances make quantum watermark image extracted have a good robustness when stego image was subjected to various geometric attacks, such as rotation and scaling. Experimental simulation based on MATLAB shows that the new algorithm has a good performance on robustness, transparency and capacity.

Zhiguo Qu, Zhenwen Cheng, Mingming Wang

Adaptive Hybrid Wavelet Regularization Method for Compressive Imaging

This paper proposes a hybrid method that simultaneously considers sparsity in wavelet domain and image self-similarity by using wavelet L1 norm, nonlocal wavelet L0 norm regularization in image compressive sensing (CS) recovery. An auxiliary variable is then introduced to decompose this composite constraint problem into two simpler regularization sub-problems. Based on Fast Iterative Shrinkage-Thresholding Algorithm (FISTA), the sub-problems corresponding to the wavelet L1 norm and the nonlocal wavelet L0 norm are then solved by soft thresholding and adaptive hard thresholding respectively. The threshold of the later is decreased according to the energy of measurement error, leading to an adaptive hybrid regularization method. Experimental results show that it outperforms several excellent CS techniques.

Lingjun Liu, Weiyu Yu, Cui Yang

Registration of OCT Fundus Images with Color Fundus Images Based on Invariant Features

Disease diagnosis and treatment are often supported by multiple images acquired from the same patient. Multimodal retinal fundus image registration techniques are fundamental to integrate the information gained from several fundus images for a comprehensive understanding. In this paper, we proposed an algorithm for registration of OCT fundus images (OFIs) with color fundus photographs (CFPs) based on invariant features. The local similarity function is defined based on the blood vessel ridges of retinal fundus images. According to the local maximum similarity function, we can extract effective image blocks and then acquire the feature matching points. We can finally achieve the registration by utilizing the quadratic surface model to calculate the transformation matrix parameters. The proposed algorithm was tested on a sample set containing 3 normal eyes and 18 eyes with age-related macular degeneration. The experiment demonstrates that the proposed method has high accuracy (root mean square error is 111.06 μm) in different qualities for both of color fundus images and OCT fundus images.

Ping Li, Qiang Chen, Wen Fan, Songtao Yuan

View-Based 3D Model Retrieval Based on Distance Learning

As information technologies develop, 3D model retrieval is paid more and more attentions by researchers. But the limitations of image features poses a great challenge to view-based 3D model retrieval. In this paper, a novel 3D model retrieval method based on distance learning is introduced. The objective function with respective to two latent variables was formulated especially. The variables are the clique information in the original graph and the pairwise clique correspondence constrained by the one-to-one matching. The proposed method has the following benefits: (1) the local and global attributes of a graph with the designed structure can be preserved; (2) redundant and noisy information can be eliminated by strengthening inliers and suppressing outliers; and (3) the difficulty of defining high-order attributes and solving hyper-graph matching can be avoided. By extensive experiments on ETH, NTU60 and MV-RED datasets with Zernike moments, Histograms of Oriented Gradients (HoG) and convolutional neural networks (CNN) features, the effectiveness of the proposed method could be tested.

Yang Shi, Nannan Liu, Xingjian Long, Lei Xu

PCANet for Color Image Classification in Various Color Spaces

Principal component analysis network (PCANet), which is a recently proposed novel deep learning algorithm, has aroused the interest of a wide variety of researchers. In this paper, we evaluate the performance of PCANet in various color spaces on different types of color image dataset. Experimental results on CURet texture database, UC Merced land use database, and Georgia Tech face database show that Luminance and Chrominance based principal component analysis network outperforms other color spaces in the vast majority of cases. Therefore, when dealing with the problem of color image dataset classification, Luminance and Chrominance based principal component network is recommended.

Jiasong Wu, Shijie Qiu, Rui Zeng, Lotfi Senhadji, Huazhong Shu

A Method of Group Behavior Analysis for Enhanced Affinity Propagation

With the popularity of mobile phones, it is necessary to mine and analyze the user habits, network applications and other data, which can help provide users with a strong adaptability of information services. On the basis of acceleration sensor and touch screen data, we analyze the behaviors of browsing the web, chatting, making calls and playing game. The traditional Affinity Propagation algorithm analyzes all the characteristics of the data as an equal role in group behavior analysis, which has some limitations. In this paper, an Adaptive Feature Weighting based on Affinity Propagation (AFWAP) Group Behavior Analysis Algorithm is proposed, which introduces feature weight into the AP algorithm. The proposed method makes different contribution to the class center in each iteration process, and assigns a new weight for each dimension attribute then to update the feature weight adaptively. In the clustering process, the importance of different features can be measured, which solves the shortcomings of the traditional AP algorithm using equal weight. Finally we apply the proposed method to group behavior analysis.

Xinning Li, Zhiping Zhou, Lele Liu

A Median Filtering Forensics Approach Based on Machine Learning

Today manipulation of digital images has become easy due to powerful computers, advanced photo-editing software and high resolution capturing devices. Verifying the integrity of images without extra prior knowledge of the image content is an important research field. Since some general post-operations, like widely used median filtering, can affect the reliability of forensic methods in various ways, it is also significant to detect them. Current image median filtering forensics algorithms mainly extract features manually. In this paper, we present a new image forgery detection method based on machine learning, which utilizes a convolutional neural networks (CNN) to automatically learn hierarchical representations from the input images. A modified CNN architecture is specifically designed to identify traces left by the manipulation. The experimental results on several public datasets show that the proposed CNN based model outperforms some state-of-the-art methods.

Bin Yang, Zhenyu Li, Weifeng Hu, Enguo Cao

Spatiotemporal Radio Tomographic Imaging with Bayesian Compressive Sensing for RSS-Based Indoor Target Localization

Wireless sensor network based device-free localization (DFL) is now widely used in security and monitoring systems for indoor and outdoor areas. Multipath fading induced noises often degrade the performance of the DFL security system. To address this problem, the paper firstly presents a spatiotemporal radio tomographic imaging (RTI) approach for the enhancement of localization. Specifically, the task of RTI can be formulated into a sparse Bayesian learning problem. In addition, two robust sparse Byesian learning algorithms are developed to handle with the low signal-to-noise-ratio (SNR) with heterogeneous noise. The proposed spatiotemporal RTI approach performs much better than traditional RTI with lower average errors in our four diverse cluttered indoor scenes. The localization results also highlight advantages of applying proposed robust sparse Bayesian learning algorithms in addressing missing estimations and outlier errors, and finally improving indoor target DFL performance.

Baolin Shang, Jiaju Tan, Xiaobing Hong, Xuemei Guo, Guoli Wang, Gonggui Liu, Shouren Xue

Detection in SAR Images Based on Histogram and Improved Elitist Genetic Fuzzy Clustering

Change detection in Synthetic Aperture Radar Images has been an important technique for Synthetic Aperture Radar Images. In this paper, a novel unsupervised change detection algorithm based on histogram and improved elitist genetic fuzzy clustering is proposed. First, a difference image is generated by multiplying transform fusion. Second, we use the characteristics of the histogram to deal with the difference image. Then, the new algorithm is proposed to partition these characteristics into changed and unchanged regions. The proposed algorithm has the following merits: 1. FCM is employed to initialize the population and to calculate the fitness function of the genetic algorithm. 2. The optimal solution is selected by an elitist selection strategy based on population concentration and the optimal solution will be the initial clustering center of FCM, which significantly increases the convergence speed. 3. The histogram is utilized to reduce the sample points of images. Compared with the state-of-the-art algorithms, the experimental results demonstrate the effectiveness in processing of change detection in SAR images.

Ronghua Shang, Weitong Zhang, Licheng Jiao

Research on Cognitive Radio Spectrum Sensing Method Based on Information Geometry

Making using of the emerging information geometry theory, we analyze the statistical properties of wireless spectrum signals received by secondary users, and propose a cognitive radio spectrum sensing method based on information geometry. We introduce a new detection structure, using the sample covariance matrix and corresponding to the points on the statistical manifold, by calculating the distance between them and make a decision, thus transforming the statistical detection problem into the geometric problem on the manifold. We also used two solutions: Constant False Alarm Rate (CFAR) Detector and Distance Detector (DD). The simulation results reveal that the performance of the information geometry method is superior to the traditional spectrum sensing algorithm, and research results will help us to explore the spectrum sensing problem from a new perspective.

Qiang Chen, Pin Wan, Yonghua Wang, Jiangfan Li, Yirui Xiao

Android Malware Detection Using Hybrid Analysis and Machine Learning Technique

This paper proposes a two-stage Android malware detection and classification mechanism based on machine learning algorithm. In this paper, we use the static analysis method to extract the software’s package features, permission features, component features and triggering mechanism. Then we use the dynamic analysis tools to obtain the dynamic behavior characters of the software, and format the static and dynamic features. Finally, we use the machine learning algorithm to deal with the feature eigenvectors in two stages, and then we will get the malicious classification of the software. The experimental results show that in the data set used in this paper the proposed method based on the combination of dynamic and static malicious code detection is more accurate than the common detection engine, and the ability of classifying malicious family is much stronger.

Fan Yang, Yi Zhuang, Jun Wang

Workload-Aware Page-Level Flash Translation Layer for NAND Flash-Based Storage Systems

Demand-based flash translation layer is an efficient page-level flash translation layer, which can effectively reduce the RAM (Random Access Memory) footprint of NAND flash-based storage systems. However, this demand-based flash translation layer does not consider the spatial locality of workloads. In this paper, a new workload-aware page-level flash translation layer is proposed for NAND flash-based storage systems. The proposed flash translation layer maintains three caches in RAM to cache mapping entries, which are the on-demand mapping entry cache, frequent mapping entry cache, and dirty mapping entry cache. Considering both temporal locality and spatial locality of workloads, the on-demand mapping entry cache is designed to store the on-demand mapping entries and sequential mapping entries. Considering the access frequency of workloads, the frequent mapping entry cache is designed to cache the most frequently accessed mapping entries. To decrease the number of updates to translation pages, the dirty mapping entry cache is designed to cache the dirty mapping entries and flush the dirty mapping entries belonging to the same translation page to NAND flash memory in a batch mode. The experimental results show that the proposed flash translation layer performs better than existing page-level flash translation layers.

Huibing Wang, Mingwei Lin, Jinbo Xiong, Li Lin, Ruliang Xiao

Biclustering Evolutionary Spatiotemporal Community in Global Trading Network

Detecting evolving communities in dynamic weighted networks are significant for understanding the evolutionary patterns of complex networks. In this paper, a novel algorithm is proposed to detect overlapping evolutionary spatiotemporal communities in the global trading network, a dynamic weighted network. This algorithm is capable of discovering those edges with similar evolving trend in a weighted community, and revealing the evolutionary of nodes and edge weight vectors simultaneously. Experiments on the global trading network show that the proposed algorithm can discover more evolving behaviors and properties which hide in those seemingly stable community structures.

Leiming Yan, Zeyu Chen, Ping Zang

Optimization and Classification

Frontmatter

Cost Optimization for Time-Bounded Request Scheduling in Geo-Distributed Datacenters

To cope with the growing service requests, a large number of cloud services are deployed in geographically distributed datacenters for better performance. Currently, how to optimize the monetary expenditure spent on VM (Virtual Machine) rental has been widely concerned. Especially, the diversities of the rental prices and service capabilities in geo-distributed regions make the problem more complicated. In this paper, the time restriction of requests and load balance are both taken into account when optimizing the VM rental cost. A two-layer geo-distributed request scheduling algorithm is presented respectively for internal and external datacenters to reduce the VM rental cost. To provide datacenter-level load balance and SLA (Service-Level Agreement) assurance, the proposed algorithm not only considers new arrival requests, but also re-dispatches requests being served to other datacenters. Finally, our work is evaluated and compared with the previous scheduling algorithms in small and large scale. Experimental results demonstrate the effectiveness of the proposed algorithm.

Xiaohui Wei, Lanxin Li, Xingwang Wang, Yuanyuan Liu

The Application of Naive Bayes Classifier in Name Disambiguation

Name repetition exists in the academic resource management system, which brings difficulties to academic evaluation, information retrieval, citation analysis and so on. According as different authors use function words in different habits, the Naive Bayes classifier was used to study in this paper. Based on the assumption of feature independence, this paper selects 26 common function words with high frequency as statistical frequency standard, use Naive Bayes classifier to classify texts. Experiments show that the method has a high accuracy rate.

Na Li, Jin Han

Mobility Prediction-Based Service Scheduling Optimization Algorithm in Cloudlets

Cloudlet is an emerging technology in mobile cloud computing. However users may be far away from cloudlets due to the mobility of mobile users, which leads to a poor network connectivity, thus, user experience will be poor. While a user moves across multiple cloudlets areas, issues of service scheduling between cloudlets to better support user experience become important. In this paper, we consider a latency-sensitive and stateful service scheduling problem in cloudlets. We propose a novel cloudlet service model and formulate the problem with the goal of finding the optimal service running sequence which minimizes the average service response time during the whole running process of the service for a user. To solve this problem, we propose an algorithm called Mobility Prediction-based Markov Decision Process (MPMDP). The proposed algorithm takes user’s mobility prediction into account, and makes an decision based on Markov Decision Process to decide on which cloudlet the service should run for a user each time. Finally, we evaluate the effectiveness of the proposed MPMDP algorithm by simulations with real-world users’ traces. The simulation result shows our algorithm achieves a lower average response time compared with previous schemes.

Lei Shi, Xi Fu, Jing Li

Single Appliance Recognition Using Statistical Features Based k-NN Classification

Recognizing the appliance according to the flowed electric current through it is quite a meaningful work which can help the electric management system to make effective policy of energy conservation. We designed an algorithm based on an improved k-nearest neighbor which can classify the unlabelled appliances’ running power data into its most similar data clusters. In other words, this algorithm is able to recognize the appliance only according to its running power data series. The classification is based upon the multifarious features extracted from the time series data sensed from the running appliance with the power metering sensors. Appliance recognition is performed with a mean accuracy over 90% in five-class classification problem.

Qi Liu, Hao Wu, Xiaodong Liu, Nigel Linge

Performance Measurement and Configuration Optimization of Virtual Machines Based on the Bayesian Network

It is significant to accurately measure the performance of virtual machines (VMs) and reasonably allocate resources according to users’ requirements for both users and cloud resource providers in IaaS cloud computing. In this paper, we propose a Bayesian network based model, called PPBN, to describe uncertain relationships among properties and performance of VMs and then measure VM performance in the form of probabilities. Further, we design a linear optimization approach to minimize resource cost and improve host resource utilization at the same time. Experimental results show that our method can measure VM performance accurately and the achieved configuration can meet users’ performance requirements well.

Jia Hao, Binbin Zhang, Kun Yue, Juan Wang, Hao Wu

Image Recapture Detection Through Residual-Based Local Descriptors and Machine Learning

At present, the tamper evidence would be invalid in recaptured image in terms of most of the digital image forensics, so the authenticity of the image detection is a security threat. Since dense local descriptors and machine learning have been successfully applied in steganalysis and forgery detection, we propose a new image recapture detection method based on these two techniques. The local descriptors were recently proposed in the field steganalysis, and some descriptors are selected by greedy strategy in the experiments. Support vector machine and ensemble classifier are utilized as the classifier in the proposed method. The experimental results show that the proposed method achieves a good performance rate that exceeds 99.61% of recaptured images and 96.40% for single captured images on the open source database.

Jian Li, Guojing Wu

A Method Towards Community Detection Based on Estimation of Distribution Algorithm

Estimation of Distribution Algorithm (EDA) is a stochastic optimization algorithm based on statistical theory. It has strong global search ability, but it is easy to fall into the local optimal solution and can not get good results in community detection. In order to solve this problem, we propose a community detection algorithm based on Estimation of Distribution Algorithm, named EDACD, whose basic framework refers EDA and the target function is modularity. EDACD keeps population diversity by adding crossover mutation operation of Genetic Algorithm as well as the improvement of probability model. Genetic Algorithm is based on “micro” level of gene, which has good local optimization ability; EDA uses the evolutionary method based on “macro” level of search space, which has strong global search ability and fast convergence speed. Taking advantage of the two methods, EDACD can used to improve the search ability of algorithm from “micro” and “macro” two levels. Finally, by experimenting on some typical real-world networks and computer-generated networks, the experimental results show that the proposed algorithm can detect the community division accurately, and has higher clustering precision compared with some representative algorithms. In addition, the proposed algorithm also has a fast convergence rate.

Yawen Chen, Wenan Tan, Yibo Pan

Sentiment Analysis with Improved Adaboost and Transfer Learning Based on Gaussian Process

Sentiment analysis is an increasingly important area in NLP to extract opinions and sentiment expressed by humans. Traditional methods are often difficult to tackle the problems of different sample distribution and domain dependence, which seriously limits the development of sentiment classification. In this paper, a novel sentiment analysis method is proposed by combining improved Adaboost and transfer learning based on Gaussian Processes to solve these two problems. A Paragraph Vector Model is employed to obtain the continuous distributed vector representations. Then, Adaboost method is used to choose the most important training features in source training data and auxiliary data. Finally, an asymmetric transfer learning classifier is introduced in Gaussian Processes. It is shown that, compared with the existing algorithms, our method is more effective for the different sample distribution and domain dependence.

Yuling  Liu, Qi Li, Guojiang Xin

HL-HAR: Hierarchical Learning Based Human Activity Recognition in Wearable Computing

In recent years there have been many successes of recognizing the human activity using the data collected from the wearable sensors. Besides, many of these applications use the data from the smartphone. But it is also a challenge in practice for two reasons. Most method can achieve a high precision in the cost of increasing memory consumption, or asking for complicated data source. In this paper, (1) Utilizing Plus-L Minus-R selection to single out the optimal combination from the feature vector extracted; (2) Introducing a fast classification method named H-ELM to resolve the problem of the highly memory consumption in the process of calculation. The main benefit of this factor is to reduce memory usage and increase recognition accuracy with a brief feature vector so that a wearable device can identify activities all by itself. And the wearable device can recognize the sample activities even if keeping away from cellphone. Our results show that this method leads to that we can recognize object activities with the overall accuracy of 93.7% in a very short period of time on the dataset of Human Activity Recognition Using Smartphones Dataset. The selected 25-dimension feature vector nearly contains all the information and after many times of test, it can achieve very high percentage of accuracy. Moreover, the method enables the learning velocity to outperform the state-of-the-art on the Human Activity Recognition domain.

Yan Liu, Wentao Zhao, Qiang Liu, Linyuan Yu, Dongxu Wang

An Improved Quantum-Inspired Evolutionary Algorithm for Knapsack Problems

As a well-known combinatorial optimization problem, knapsack problems commonly arise in security areas. In this paper, an improved quantum-inspired evolutionary algorithm (PEQIEA) is proposed to solve knapsack problems. In PEQIEA, in each iteration, the state preference of the elite group is used to update the group. The elite group of each iteration consists of a certain number of individuals which are selected by their fitness values. A state preference is proposed to improve the efficiency of the algorithm. A new quantum-inspired gate is obtained by the elite group and their state preference. The Q-gate is then used to make the evolution of the group. The parameters in PEQIEA, which affect the accuracy and efficiency of the algorithm, are discussed empirically. The performance of PEQIEA is then evaluated through extensive experiments.

Sheng Xiang, Yigang He, Liuchen Chang, Kehan Wu, Chaolong Zhang

Output-Based Sampled Data Control for Linear Systems: A Measurement Error Estimation Approach

In this paper, the output based sampled data control problem for linear systems is investigated. First, based on the measured output, a closed-loop system with measurement error has been obtained. Then, in order to obtain the maximum upper bound of the sampling interval, a measurement error model has been derived to estimate the measurement error. By using the Gronwall-Lemma and the Lyapunov stability theory, the control gain matrix and the maximum upper bound of the sampling interval has been explicitly given.

Yuelin Shen, Wenbing Zhang

Adaptive Firefly Algorithm with a Modified Attractiveness Strategy

The performance of firefly algorithm (FA) is seriously affected by its parameters. Recently, we proposed a new FA with adaptive control parameters (ApFA), in which the step factor is dynamically updated and the attractiveness oscillates in a fixed interval. In this paper, we present a modified version of ApFA, namely MApFA, which introduces a new strategy to change the attractiveness. Simulation results on several benchmark functions show that MApFA can achieve more accurate solution than ApFA.

Wenjun Wang, Hui Wang, Jia Zhao, Li Lv

Newton Method for Interval Predictor Model with Sphere Parameter Set

In this paper, we study the construction of interval prediction model. After introducing the family of models and some basic information, we present the computational results for the construction of interval predictor models, using linear regression structures which regression parameters are included in a sphere parameter set. Given a size measure to scale the average amplitude of the predictor interval, one optimal model that minimizes a size measure is efficiently computed by solving a linear programming problem, firstly we apply the active set approach to solve the linear programming problem and propose one Newton iterative form of the optimization variables. Based on these optimization variables, the predictor interval of the considered model with sphere parameter set can be directly constructed. Secondly as for a fixed non-negative number from the size measure, we propose a better choice by using the Karush-Kuhn-Tucker optimality conditions.

Xuan Xiao, Peng Wang, Jian-Hong Wang

A Conjugate Gradient Algorithm with Yuan-Wei-Lu Line Search

This paper presents a three term conjugate gradient algorithm and it has the following properties: (i) the sufficient descent property is satisfied; (ii) the algorithm has the global convergence for nonconvex functions; (iii) the numerical results are more effective than that of the normal algorithm.

Gonglin Yuan, Wujie Hu, Zhou Sheng

IU-PMF: Probabilistic Matrix Factorization Model Fused with Item Similarity and User Similarity

Probabilistic Matrix Factorization has been proven a very successful model for recommending because of scalability, accuracy and the ability to handle sparsity problem. However, many studies have demonstrated that PMF alone is poor to reveal local relationships which can be captured by neighborhood-aware methods. In this paper we present the IU-PMF model fusing Item Similarity and User Similarity in PMF, which combines the merits of both methods. The IU-PMF model consists of two phases: the Item and User similarity matrices computation phase not needing to be applied frequently; the fused PMF model solving phase which scales linearly with the number of observations. The IU-PMF model incorporates Item similarities and User similarities abstracted from User-Item ratings into the PMF model, which helps to overcome the often encountered problem of data sparsity, scalability and prediction quality. Experiments on three real-world datasets and the complexity analysis show that IU-PMF is scalable and outperforms several state-of-the-art methods.

Yilong Shi, Hong Lin, Yuqiang Li

New Delay-Dependent Stability for Neutral Systems with Its Application to Partial Circuit Model

The issue on robust stability for a class of uncertain linear neutral systems with time-varying delays is studied. Together with multiple integral functional technique and using some novel integral inequalities, the much tighter estimation on derivative of Lyapunov functional is presented and one stability criterion is presented in terms of linear matrix inequalities (LMIs), in which those previously ignored information can be reconsidered. Especially, the multiple Lyapunov functional terms include the interconnection between neutral delay and state one. Finally, some comparing results with application to partial element circuit model can show the benefits of our conditions.

Tao Li, Ting Wang, Jin Deng, Li Zhang

A Quantitative Evaluation Method of Surveillance Coverage of UAVs Swarm

Small unmanned aerial vehicles (UAVs) have native advantages in wide area surveillance. Pursuing the spreading of the UAV swarm as dispersive as possible is an effective method of improve the performance of surveillance. In this paper, quality of deployment issue is surveyed and analyzed in term of a novel measure - deployment entropy. The idea of deployment entropy comes from Shannon’s information entropy. Deployment entropy could help operators to obtain the whole understanding of the interested region from describing the circumstances of every sub region. The more dispersive UAVs are deployed, the greater the deployment entropy we can get. From numerical simulation, results show that by computing the value of deployment entropy, it is possible to evaluate the distribution of UAVs in a wide area, and the burden of calculation is less than traditional evaluation method.

Wei Li, Changxin Huang, Kaimin Chen, Songchen Han

Chaos Prediction of Fast Fading Channel of Multi-rates Digital Modulation Using Support Vector Machines

According to the support vector domain properties, the paper establishes vector domain predictive models of chaos channel as well as chaos phase trace of non-linear map, the chaotic fading channel model was established based on Takens phase space delay reconstructing theory. Self-learning makes error least upper bound of generalization model to be minimum. The non-linear higher dimension map was realized by the squares support vector domain. The future fading channel data was predicted from training data set. The predictive error changes with the increase of embed dimension to a constant. The experiment result indicates that the support vector domain needs little support vector with fast convergence rate. With the small sample and unknown probability density, the multi-path predictive series consisted with true value series in Doppler fast fading channel. Under the conditions of small sample, the predicted series is in concordance with the channel true value.

Yijing Ren, Ren Ren

Short Paper

Frontmatter

System Log-Based Android Root State Detection

Android rooting enables device owners to freely customize their own devices. However, rooting system weakens the security of Android devices and opens the backdoor for malware to obtain privileged access easily. For this reason, some developers have introduced detection mechanisms for sensitive or high-value mobile apps to mitigate the potential security risks. Nevertheless, the existing root prevention and detection methods generally lack universality. In this paper, we studied the existing Android root detection methods and found the both parties have ignored the traces of the relevant behavior in the log. Thus, we proposed the system log based root state detection method. In the method, we directly use the existing log information to find clues to verify the system root state on one hand, on the other hand, to use the triggering features of some special operations to update and enrich the log information. The results show that, even be deliberately erased, some log information is still remained which can be used to verify whether system was rooted or not.

Junjie Jin, Wei Zhang

Quantified Attribute Access Control Model for Cloud Storage Platform

Recently, cloud computing is the most domain studies in information technology. At the same time, the security of cloud computing becomes an important challenge. Existing access control models are poor on the granularity of the model elements and the dynamics which leads the security of the resource in cloud computing is limited. In this paper, a quantified attribute-based access control (QABAC) model is proposed. The concept of quantified attribute and trust degree is defined. Three attribute quantization functions are proposed for dynamically calculation, and the security degree of the access will be obtained. Finally the authorization policy determines final permission according to the trust degree. Compared with other traditional models, QABAC is flexible, extensible and dynamic. It will not only protect the security of resource among potential attack from network, but also has the capacity to meet the performance requirement in practical applications.

DongMin Li, Jing Li, Sai Liu, Chao Wang

Dynamically-Enabled Defense Effectiveness Evaluation in Home Internet Based on Vulnerability Analysis

Current intelligent devices in Home Internet, such as routers and cameras, have suffered malicious attacks from hackers. Therefore, security for Home Internet appears particularly significant. In order to have a quantitative evaluation of security defense ability of Home Internet system, this paper proposes an improved vulnerability scoring method on Home Internet based on Information Security Technology Security Vulnerability Classification Guide. Compared to original scoring method which is mainly based on Internet, this improved scoring performs differently. It’s aimed to have a quantitative evaluation on security defense effectiveness of Home Internet system: higher vulnerability score indicates higher threaten degree and relatively weak defense ability. In this paper, the Home Internet system takes dynamically-enabled defense technology (randomly changes system status) to make defense. Through calculating vulnerability scores before and after random changes of system status, this paper succeeds in making a quantitative evaluation on security defense ability of Home Internet system.

Ting Wang, Min Lei, Jingjie Chen, Shiqi Deng, Yu Yang

The Formal Transformation of AADL Based on Z-CoIA

The Architecture Analysis and Design Language (AADL) is a component-based semi-formal language. This paper proposes an expanded component-interaction automaton with Z language (Z-CoIA) based on the characteristics of AADL, introducing the formal specification Language Z into the component-interaction automata, then the formal transformation rules from AADL to the Z-CoIA is given, which is good for describing the data during system interaction and the attributes in state transitions and data constraints. Finally, a concrete example is shown.

Fugao Zhang, Zining Cao

Masking Signature Data Errors of Software-Based Control Flow Checking Techniques Employing Redundancy Signature

With technology scaling, transient faults are becoming an increasing threat to hardware reliability. Commodity systems must be made resilient to these in-field faults through very low-cost resiliency solutions. Up to 77% of the transient faults cause Control Flow Errors (CFEs). Software-based control-flow checking techniques have emerged as promising low-cost and effective solutions. The signature monitoring method is the foundation of most of these control flow checking techniques. Some CFEs cannot be detected by previous control flow checking techniques when transient fault hit the software signature. A technique, masking signature data errors of software-based control flow checking techniques employing redundancy signature (CFCRS), with the ability to mask these CFEs is proposed in this paper. In CFCRS, these errors can be detected and corrected by triple redundancy signature. The experimental results demonstrated that CFCRS is able to mask all 2,000 injected faults in software signatures; It is reasonable and feasible to apply this technique on the former software-based control flow checking techniques due to its perfect correction coverage of CFEs caused by incorrect-signature and low overheads.

Liping Liu, Linlin Ci, Wei Liu

The Analysis of Key Nodes in Complex Social Networks

Key nodes play really important roles in the complex socail networks. It’s worthy of analysis on them so that the social network is more intelligible. After analyzing several classic algorithms such as degree centrality, betweenness centrality, PageRank and so forth, there indeed exist some deficiencies such as ignorance of edge weights, less consideration on topology and high time complexity in the research on this area. This paper makes three contributions to address these problems. Firstly, a new idea, divide and conquer, is introduced to analyze directed-weighted social networks in different scales. Secondly, the improved degree centrality algorithm is proposed to analyze small-scale social networks. Thirdly, an algorithm named NodeRank is proposed to address large-scale social networks based on PageRank. Subsequently, the effectiveness and feasibility of these two algorithms are demonstrated respectively with case and theory. Finally, two representative basesets with respect to the social networks are adopted to mine key nodes in contrast to other algorithms. And experiment results show that the algorithms presented in this paper can preferably mine key nodes in directed-weighted complex social networks.

Yibo Pan, Wenan Tan, Yawen Chen

Novel Schemes for Bike-Share Service Authentication Using Aesthetic QR Code and Color Visual Cryptography

To enhance bike-share service authentication, we propose two novel schemes based on aesthetic QR code combining XOR-based visual cryptography scheme (XVCS). In Scheme I, aesthetic QR code based on error correction mechanism and XVCS are combined for authentication. In comparison, Scheme II exploits aesthetic QR code based on XOR mechanism of RS with Positive Basis Vector Matrix (PBVM) and XVCS. The larger region of secret information and the better visual appearance of aesthetic QR code are shown in experiments.

Li Li, Lina Li, Shanqing Zhang, Zaorang Yang, Jianfeng Lu, Chin-Chen Chang

An Improved RFID Search Protocol

In this paper, we propose an improved scheme to solve the security risks of S-S’s scheme. In order to resist the replay attack, the improved protocol uses the pseudo identifier XOR to encrypt the random number generated by the reader; the random number generated by the reader is added to the tag’s response message to resist the tag impersonation attack. In addition, this paper uses Avoine model to analyze the privacy of the improved protocol. The theoretical analysis shows that the proposed scheme can effectively resist against replay attack, tag impersonation attack, and de-synchronization attack. Moreover, the improved scheme can provide forward untraceability and tag untraceability. Compared with the existing RFID tag search protocol, the computational complexity of the tag and gate complexity on the tag side in the improved protocol is lower, the number of interaction with the reader is less, so the search for low-cost tags can be implemented more efficiently.

Ping Wang, Zhiping Zhou

A Developmental Evolutionary Algorithm for 0-1 Knapsack Problem

In this paper, a developmental evolutionary algorithm (DEA) is proposed, which mainly based on the developmental evolutionary and learning theory. We regarded the chromosome individual that in EC as an autonomous development individual; and developed mental capabilities through autonomous real-time interactions with its environments by using development learning methods under the control of its intrinsic developmental program, when chromosome individual achieved the development objective, genetic operation started immediately, otherwise continue developing. Finally, we used DEA to solve the 0/1 knapsack problem and designed experiment to compare with QEA, ACO. Experimental results showed that DEA has better convergence, and can effectively avoid falling into local optimal solution.

Ming Zhong, Bo Xu

WiSmart : Robust Human Access and Identification for Smart Homes Using WiFi Signals

In smart homes environment, the access and identification of different persons are key enabling technology to make personalized services or intrusion detection. However, most state-of-the-art systems require users to carry or wear dedicated devices which are not user-friendly. In this paper, we present WiSmart, a fine-grained device-free framework that can distinguish different actions and identify persons within a short duration using WiFi signal. The experiments show that we can achieve an average accuracy of human activity model up to 89.14% and human identification model greater than 85.6%.

Shangqing Liu, Yanchao Zhao, Bing Chen

Robust Face Recognition Model with Adaptive Correction Term via Generalized Alternating Direction Method of Multipliers

During the past few years, face recognition technique has received significant attention in the fields of computer vision, neuroscience, psychology and others. The robust face recognition casts the problem as an $$\ell _1$$ℓ1-minimization problem to find a sparse representation of the test image in terms of the training set. The main purpose of this paper is to firstly construct an $$\ell _1$$ℓ1-$$\ell _1$$ℓ1-minimization model and secondly be solved via a generalized alternating direction method of multipliers. Most importantly, the model proposed therein contains an adaptive correction term to get sparse representation with higher accuracy. Extensive experiments on the simulated data verify that the proposed method is effective.

Can Wu, Yunhai Xiao, Wen-Jie Liu

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise