Advances in Intelligent Information Hiding and Multimedia Signal Processing

In recent years, in addition to one-dimensional reversible data hiding many two-dimensional histogram modification methods have been proposed. In this paper, various two-dimensional histogram modification reversible data hiding method are studied, analyzed, and compared, providing useful information for researchers in this field. The analysis and comparison of these include two-dimensional histogram modification reversible data hiding methods based on: difference-pairs, pairwise prediction error expansion, and PVO and pairwise prediction error expansion.

In 2010, Chen et al. proposed a steganography mechanism using a hybrid edge detector. The hybrid edge detector is combining the Canny and the fuzzy edge detectors. However, although this method has a high embedding payload (2.10 bpp) it is an irreversible data hiding scheme. A considerable amount of research has since been conducted to provide an improved method of data hiding that removes LSBs of the image pixel value and performs edge detection. This increases the payload as it is not necessary to store edge information in the stego image. This paper conducts a comprehensive study of irreversible and reversible data hiding based on the image texture data hiding method, and compares embedding capacities and visual image qualities.

The Pixel-Value-Differencing (PVD) method is a popular irreversible data-hiding scheme proposed by Wu and Tsai (2003). The design of the interval range plays an important role in this method. It is intuitive to design the interval range using the width of the power of two, and the number of bits that can be embedded in a block of two consecutive pixels is determined based on the width of the interval range to which the difference value belongs. In 2013, Tseng and Leng broke the rules in that they designed the interval range based on a perfect square and achieved a higher payload and lower distortion. Based on Tseng and Leng’s idea, we propose a new data-hiding strategy based on the PVD method for an optimal design of the interval range.

In 2006, Zhang and Wang proposed a novel data hiding method, namely the exploiting modification direction (EMD) scheme. In their scheme, the secret message in a $$ (2n + 1) $$(2n+1)-ary notational system is carried by n cover pixels, where n is a system parameter, and at most one pixel is increased or decreased by 1. The EMD scheme exhibits two disadvantages: the secret message must be converted into a non-binary system, and the embedding capacity is limited. In this study, a new data hiding scheme, based on a regular octagon-shaped shell, is proposed to overcome the EMD scheme’s shortcomings.

On the Internet, various communication channels are readily available for companies, communities, and individuals. As companies have increasingly made use of social websites and instant messaging technologies for marketing and public relations, additional effort has been put in place to create interactive web contents to attract online audience. It is easy for anyone to simply copy and modify existing web contents for their own use. A common approach to protecting the copyright of a companys web content is digital watermarking with copyright information. The watermarking technique is a good way to achieve the goal of copyright protection. This paper presents a web page copyright protection method by integrating Cartesian product combination, CSS and HTML tag capitalization method, HTML attribute combination method, HTML attribute quotation mark method, and CSS attribute value embedding method. The experimental results demonstrate that the proposed method success achieved the goal of copyright protection for web pages.

In order to help individuals effectively manage and record these health check physiological measurement data, this research developed an “Integrated Health Check Report Analysis and Tracking Platform” together with H&B Health Centers. Using this platform, the public can query health check data and analysis charts from health check centers through a website. The information will include suggestions from doctors, nutritionists and representatives of various health check categories. The meanings behind various data can be explained to the general public using illustrations.The “Integrated Health Check Report Analysis and Tracking Platform” can retain health check data from recent years and the system will provide sharing functions so that friends and family are also able to care for the patient. The system will also target various abnormal test data and provide nutritional recommendations in the specific category. This will prevent the patient from using the wrong medicine or remedies, resulting in more serious consequences. Using the health check data from the complete history of records from the platform, it can provide analysis and trend graphs on various data to allow laypersons to understand their own health conditions and trends in simple ways via graphics. In addition, using H&B Health Center’s big data combined with personal lifestyle assessment and health check data over the years will allow automatic generation of routine health check recommendations and reminders for regular checks.

Since 1997, internet-based e-government has been promoted by the Research and Development Review Board of Executive Yuan, R.O.C., and the application of electronic forms is one of the e-government processes. Electronic form system includes integrated document management and files management systems and is integrated with an electronic attendance system as well. In this way, other than merely achieving creating a paperless system, it also reduces the manpower required for paper-form transmission and shortens document delivery time.This study took Tainan city government and its subordinate organization as the study objects. Owing to the merging of Tainan County and Tainan City a couple of years ago, their old electronic attendance systems were redesigned and reconstructed into a single system. Through this example, this study investigated how to design an electronic attendance system which could integrate various business characteristics.

The development of network and mobile technologies contributed to the emergence of mobile and ubiquitous learning. Therefore mobile and ubiquitous learning are attracting both academic and public interest in the recent year. Owing to mobile learning is defined as “the processes of coming to know through conversations across multiple contexts among people and personal interactive technologies”, this study conducts a content analysis to view the relative researches that apply this technology to facilitate various activities of learning. Those activations of learning can be designed in formal and informal and personalized learning environments during classroom lectures or outside of the classroom. The advantage of portable technologies can help learners to connect various learning activities both in formal and informal personalized learning environments.Generally, previous research in mobile learning are focus on the lower or younger learners, the broad application of mobile and ubiquitous learning in higher education settings is limited. Further the mobile learning design can be design and allow learner selects and transforms information, constructs hypotheses, and go beyond the information given. Therefore this research try to analyze the mobile learning research that are focus on higher education setting and based on constructivism theory.

Owing to there are still has difficulties of lacking a convenient way to manage various devices and provide customize service either by an appliance or recreational equipment. Therefore, it is necessary to develop a system (or device) to control and management this problem. For economic considerations, it is inevitable to have a new system (or device) for this requirement. Preferably, it can be achieved by using available or easily made hardware with properly designed software architecture for control. In present, it is made by customizing circuits with only one specification for specific equipment. It is still lacking a integrate system or device can provide and integrate and manage those various respond or feedback data. To achieve this goal, this research proposes a device with control programming system also defined data in a device database and an instruction database to solve this problem.

Air pollution has become a major threat today and the related respiration illnesses have deteriorated the wellness of many people. Herb, as one option of the medicine and also as diet, can provide very promising solution to the above threats. However, even though herbs are easily accessible in every cultures and areas, the knowledge of applying herbs on improving health is complicated and it takes quite lots of efforts to acquire the knowledge. In this paper, we design an ontology-based herb therapy recommendation web for respiration system health.

This paper develops a robust optimal control technology for multimedia signal processing (MSP) applications. The proposed control technology combines the advantages of nonsingular finite-time convergence sliding mode control (NFTCSMC) and genetic algorithm (GA). The NFTCSMC has finite system state convergence time including nonsingular merit unlike infinite-time exponential convergence of classic sliding mode control, but the chattering still occur under a highly uncertain disturbance. For multimedia signal processing applications, the chattering causes high voltage harmonics and inaccurate tracking control. To enhance the performance of multimedia signal processing, the GA is well adopted to tune the control gains of the NFTCSMC so that the chattering can be removed. Experimental results are given to conform that the proposed control technology can lead to high-quality AC output voltage and fast transient response. Because the proposed control technology is easier to realize than prior technologies and gives high tracking accuracy and low computational complexity algorithm, this paper will be of interest to designer of related MSP applications.

The paper reports the implementation of an eBook reader with emotion sensing feature through Kinect sensor device. Recently, more and more traditional paper books have been transformed into an electronic form which is so called the electronic book (eBook). Currently, eBook contents could be in the form of combining various media through visual and audio communications. However, it still lacks the capability of real interactions between human and eBook to respond to user’s emotion as a person is reading eBooks. Therefore, we will apply emotional sensing technology to enhance the human-eBook interactions. So that an eBook reader device can sense a person’s reaction that reflects the situation and meaning as described in eBook contents. In this paper, the implementation of a novel eBook reader system integrating a Kinect sensor and the associate system development kit (SDK) will be reported. We will use a widely used face recognition technology, Active Appearance Model (AAM), for detecting user’s facial expression and emotions. With such a new feature, users may have funs and more realistic interactions with eBooks as a person is reading eBooks.

Massive open online courses (MOOCs) offer valuable opportunities for freedom in learning; however, many learners face cognitive overload and conceptual and navigational disorientation. In this study, we used handouts to automatically build domain-specific knowledge maps for MOOCs. We considered handouts as conceptual models created by teachers, and we performed text mining to extract keywords from MOOC handouts. Each knowlege map is based on the structure of the handouts, each consisting of an outline, title, and content. The findings suggest that using handouts to build knowledge maps is feasible.

This paper proposes an innovated way to compose music automatically according to the input of the heartbeat sensor, to generate music with the correspondent emotion states. The typical 2D emotion plane with arousal and valence (A-V) states are adapted into our system, to determine the generative music features. Algorithmic composition technique including Markov chain is used, with the emotion - music feature mapping method, to compose the desired correspondent music. The result show a pretty good success with various generative music, including sad, happy, joyful, and angry, and the heartbeat values show its good consistency for the correspondent emotion states finally.

Regarding conventional driving scenarios, vehicles are isolated and self-secured without interaction although many vehicles may exist in a surrounding. Driving safety is thus secured by individual driving behavior; that says, drivers must take care of themselves. Whereas many technologies contribute to the rapid development of vehicular networking and applications, it can lead to better quality of driving experience. This paper proposes a system and method for inter-vehicle media content distribution for driving safety. The system aims to expand the range of a drivers’ vision by sharing front-end views of neighbor vehicles on the road. In this system, we use the Wi-Fi facility to establish an inter-vehicle communication network, use the UPnP protocol to manage vehicles, and then use RTSP to deliver real-time video to other vehicles in a network. Thus, drivers can catch a broad view through the sharing services of front views. In addition, this system provides functions to record images, road features and geographical information, and then to transfer them onto a back-end database. The system will be able to notify traffic accidents and situations on the warming map, so the drivers can browse the information to recognize what happens in front of them before approaching to the area. Accordingly, we design a system architecture, develop a software prototype, and demonstrate practical scenarios.

The authors present an economic way to hide patient diagnosis information in electrocardiogram (ECG) signal based on smart offset coefficient. Simulations indicated that hiding capacity is larger than existing techniques while the perceived quality is good. Moreover, the method is tolerant of the attacks such as inversion, translation, truncation, and Gaussian noise-addition attacks, which is rare in conventional ECG steganographic schemes. Since the privacy (or medical message) of the patients can be fast and effective embedded in ECG host by the proposed method, it is feasible for our method being employed in real-time applications.

In this paper, we propose an automatic people counting system by using our low-cost Internet-of-Things (IoT) platform consisting of a single camera and Raspberry Pi. In this system, we count the number of moving people in bidirection by observing from a side view. Because the system can determine the height of the people, our system can be used to classify them into adults or children. This system is applied for no people overlapping problem in indoor environment only. The background subtraction and morphological operations are used to extract foreground objects from background images. The experimental results show proposed method can achieve 98% of people counting accuracy. It can also achieve 91% accuracy in adult/child classification. Although the algorithms for the people counting and classification are not novel, our technical contribution is that we have implemented them onto our IoT platform, whose cost is less than 100 US dollars. In addition, the images do not need be sent to the server, but all the image processing is done inside the device and only the results are uploaded to the server. This system can be applied to for customer behavior analysis or security.

One of the most challenging issues when operating a high-power light-emitting diode (LED) is to conduct appropriate packaging materials for the reliable thermal management of the entire device. Generally, the considerable amount of heat produced around the junction area of the LED would transfers to the entire device, and causes thermal expansion in the packaging material. It induces strain inevitably, and hinders the output performances and possible applications of high-power LEDs. The coefficient of thermal expansion (CTE) is a physical quantity that indicates the expansion to which value a material will be upon heating. Therefore, as far as an advancement of thermal management is concerned, the quantitative and real-time determination of CTE of packaging materials becomes more important than ever since the demanding of high-power LEDs is increased in recent years. In this study, we measure the CTE of GaN-based (λ = 450 nm) high-power LED encapsulated with polystyrene resin by using optical coherent tomography (OCT). The displacement change between individual junctions of OCT image is directly observed and recorded to derive the CTE values of composed components of the LED device. The obtained instant CTE of polystyrene resin is estimated to be around 10 × 10−5/°C, which is a spatial average value over the OCT scanning area of 10 μm × 10 μm. The OCT provides an alternative way to determine a real-time, non-destructive, and spatially resolved CTE values of the LED device, and that shows essential advantage over the typical CTE measurement techniques.

In this paper, we study reversible data hiding schemes for JPEG images and we propose a method which is constructed by embedding secret data into zero quantized coefficients. In the proposed method, consecutive zero quantized coefficients in a fixed set of entries of DCT block are used to embed data in the following way. First, the algorithm selects several positions from this fixed set according to the first part of the secret message and modifies selected zero coefficients by small non-zero values according to the second part of the secret message. Experimental results show that our proposed methods have higher embedding capacity and obtain lower ratios between the increased file size and payload than previous methods.

Biomedical materials have different optical properties (e.g. absorption) for different wavelengths. Therefore, probing biomedical materials with multiple wavelengths not only can get in-depth understanding of the detected biomedical materials but also can differentiate the detected biomedical materials. To achieve this purpose, in this conference paper, we present our initial results of building up a portable multiple-wavelength biomedical sensing system. At this initial phase, we assembled this kind of system with multiple wavelengths of light sources and photodetectors and preliminarily tested the absorbance of the glucose solutions with different concentrations. The result shows good linearity of the absorbance of the glucose solution with concentration. In addition, we also measured the absorbance of the glucose solutions using a broadband white light source and a spectrometer. This result also exhibits linearity but different slop of absorbance with glucose concentration, which confirms the linearity result obtained from the built sensing system. The difference of slop maybe relates to the difference of optical design between these two systems.

Electronic stethoscope system is the most frequently used approach to collect respiratory sound to evaluate the lung function of patients or investigate various kinds of lung diseases of in recent years. However, current electronic stethoscope systems are just suitable for measuring respiratory sounds under static state. Under motion, the respiratory sounds are easily affected by the vibration of the human body. Moreover, it is also inconvenient to wear the conventional electronic stethoscope system. In order to improve the above issues, a novel wireless respiratory sound recording system was proposed to collect respiratory sound under motion. Here, a wireless and wearable respiratory sound recording device was designed to collect respiratory sound wirelessly. It is also easy to wear and monitor respiratory sound under motion due to its advantages of small volume and wireless transmission. Moreover, the technique of adaptive filter was also applied to enhance the noisy respiratory sound from single channel trial. From the experimental results, the noisy respiratory sound can be effectively improved by the proposed adaptive filter. Therefore, the proposed system exactly contains the potential of being a good assisting tool for lung diseases and may be applied in the applications of lung and sports medicine in the future.

This paper demonstrated a general algorithm to carry out precision alignment with UVW or XXY and XYθ stages using two cameras without rotation center information. Image processing with template match was implemented to search for fiducial mark under the field of view of a camera as the positioning tool. We established the relationship between cameras and motion stage by calibration procedure with inverse regression method. We finished the substrate’s alignment parallel to the mechanical axis in 2–4 times iterations and applied this algorithm in sixth-generation LCD prober system successfully.

Biomedical materials have different optical properties (e.g. absorption) for different wavelengths. Therefore, probing biomedical materials with multiple wavelengths not only can get in-depth understanding of the detected biomedical materials but also can differentiate the detected biomedical materials. To achieve this purpose, in this conference paper, we present our initial results of building up a portable multiple-wavelength biomedical sensing system. At this initial phase, we assembled this kind of system with multiple wavelengths of light sources and photodetectors and preliminarily tested the absorbance of the glucose solutions with different concentrations. The result shows good linearity of the absorbance of the glucose solution with concentration. In addition, we also measured the absorbance of the glucose solutions using a broadband white light source and a spectrometer. This result also exhibits linearity but different slop of absorbance with glucose concentration, which confirms the linearity result obtained from the built sensing system. The difference of slop maybe relates to the difference of optical design between these two systems.

Public key encryption with keyword search (PEKS) is one of searchable encryption mechanisms. It not only provides user to retrieve ciphertext by keyword but also protects the confidentiality of keyword. In the past, many PEKS schemes based on different cryptosystems were proposed. Recently. Zheng et al. proposed a certificateless based PEKS scheme called CLKS. In this paper, we show that Zheng et al.’s CLKS scheme has some security flaw, i.e. their scheme suffered from an off-line keyword guessing attack.

In this paper, we propose an efficient algorithm to discover HAUIs based on the compact average-utility list structure. A tighter upper-bound model is used to instead of the traditional auub model used in HAUIM to lower the upper-bound value. Three pruning strategies are also respectively developed to facilitate mining performance of HAUIM. Experiments show that the proposed algorithm outperforms the state-of-the-art HAUIM-MMAU algorithm in terms of runtime and memory usage.

Anonymous authentication allows one entity to be authenticated by the other without revealing the identity information. In mobile networks, mobile devices communicate with each other to exchange resources. To achieve anonymous mutual authentication, the devices are anonymously authenticated under the trusted server. Recently, Chung et al. proposed a efficient anonymous mutual authentication scheme for inter-device communication using only low-cost functions, such as hash functions and exclusive-or operations. However, we find that their protocol does not preserve user’s privacy in terms of untraceability. Also, their protocol is vulnerable to a denial of service attack and a user impersonation attack.

This paper presents a lossless data hiding scheme for concealing message in each block of discrete cosine transform (DCT) coefficients. Our proposed improves the Chang et al.’s scheme for obtaining higher performance in terms of embedding capacity and image quality. In our scheme, two neighbor coefficients which are near to zero as an group are selected in each block, and then, hiding the message into each group. The coefficient is modified by one unit; therefore, the stego-image can keep the high image quality. Experimental results show that the proposed scheme not only provides better performance than that of previous works but also achieves coefficients reversibility.

In this study, we focus on analysing the relationship between the foreign exchange rate and the international tourism flow. Three foreign exchange rate forecasting models including GARCH(1,1), EGARCH(1,1), and the IABC forecasting model based on the computational intelligence are employed to produce the forecasting results. The Mean Absolute Percentage Error (MAPE) is selected to be the evaluation criterion for comparing the forecasting results of these models. The experiments contain the USD/NTD foreign exchange rate and the inbound international tourism flows in years of 2009 to 2010. The experimental results reveal that adding the international tourism flow as the new reference in the forecasting process has the positive contribution to the foreign exchange rate forecasting results.

This paper investigates a kind of bilevel programming with fuzzy random variable coefficients in both objective functions and the right hand side of constraints. On the basis of the notion of Er-expected value of fuzzy random variable, the upper and lower level objective functions can be replaced with their corresponding Er-expected values. In terms of probability over defuzzified operator, fuzzy stochastic constraints can be converted into the equivalent forms. Based on these, the fuzzy random bilevel programming problem can be transformed into its deterministic one. Then we suggest differential evolution algorithm to solve the final crisp problem. Finally, a numerical example is given to illustrate the proposed method.

Decomposition-based evolutionary algorithms are promising for handling many objective optimization problems with more than three objectives in the past decade. In the proposed algorithm, we develop a new dominance relation based on - efficiency order dominance (MOEA/D-$$ \upvarepsilon{\text{EOD}} $$εEOD) in each sub-problem to realize the selection and update of the individuals. Besides, a dynamic adaptive weight vector generation method is proposed, which is able to dynamically adjust the weight vector setting according to the current distribution of the non-dominated solution set. The proposed algorithm has been tested extensively on six widely used benchmark problems, and an extensive comparison indicates that the proposed algorithm offers competitive advantages in convergence and diversity.

Multi-Objective Evolutionary Algorithm (MOEA) is emerging as a state-of-the-art methodology to solve the ontology meta-matching problem. However, the huge search scale of large scale ontology matching problem stops MOEA based ontology matching technology from correctly and completely identifying the semantic correspondences. To this end, in this paper, a large scale multi-objective ontology matching framework is proposed, which works with three sequential steps: (1) partition the large scale ontologies into similar ontology segment pairs; (2) utilize MOEA to match the similar ontology segments in parallel; (3) select the representative ontology segment alignments, which are further aggregated to obtain the final ontology alignment. In addition, a novel multi-objective model is also constructed for ontology matching problem and the MOEA and entity similarity measure that could be used in this framework are also recommended. The experimental result shows the effectiveness of our proposal.

For global optimization problems with a large number of local optimal solutions, evolutionary algorithms are efficient parallel algorithms, but they drops into local optimum easily, therefore their efficiency and effectiveness will be much reduced. In this paper, first, a new deleting strategy is proposed that can eliminate all local optimal solutions no better than this obtained local optimal solution. Second, when algorithm drops into a local optimal solution, a new jumping strategy is proposed that can jump out of the current local optimal solution and then find a better local optimal solution. Based on the above, a new algorithm called evolutionary algorithm with deleting and jumping strategies (briefly, EADJ) is proposed, and the algorithm convergence is proved theoretically. The simulations are made on 25 standard benchmark problems, and the results indicate the proposed deleting strategy and jumping strategy are effective; further, the proposed algorithm is compared with some well performed existing algorithms, and the results indicate the proposed algorithm EADJ is more effective and efficient.

The water temperature plays a major role in application of Water-Source Heat Pump (WSHP) system. The water temperature is directly related to the energy efficiency of heat pump unit and economy of WSHP system. As the air temperature is the most influential factor in the water temperature in all meteorological parameters, the effect of air temperature on the changing water temperature should be considered first when estimating the river water temperature. In this paper, we dedicate to study the changing rule of water temperature with the air temperature by analyzing over one year water temperature data. Particularly, the least square method is used for the regression analysis of 1st-order to 6th-order curve, and a fitting curve is educed, based on Cubic model of the air temperature and the average of the corresponding water temperature. The 88% of the difference values between average water temperatures and fitted water temperatures are in the range of −1.5–1.5 °C. Therefore, it can provide a simple method to determine the daily average water temperature through analyzing the relationship between the water average temperature and air temperature by using the method of Cubic fitting. The study in this paper can be used as a reference for the design and analysis of WSHP system.

Forecasting is an important technique in many industries and business fields for reading the terrain. The category of technology industry stock, which includes 7 independent stocks, in Taiwan Stock Exchange (TWSE) is selected to be the study subject in this paper. The goal is to forecast the return index of the individual stocks base on the information observed from the trading historical da-ta of the subjects. By including the trading volume, the number of trading rec-ords, the opening price, and the closing price in the inputs to the representative models in time-series and computational intelligence: EGARCH(1,1) and the In-teractive Artificial Bee Colony (IABC), respectively, the forecasting accuracy are compared by the Mean Absolute Percentage Error (MAPE) value. The experi-mental results indicate that the IABC forecasting model with the selected input variables presents superior results than the EGARCH(1,1).

The chest computed tomography (CT) is the most commonly used imaging technique for the inspection of lung lesions. In order to provide the physician more valuable preoperative opinions, a powerful computer-aided diagnostic (CAD) system is indispensable. In this paper, we aim to develop an ant colony optimization (ACO-based) classifier to extract the lung mass. We could calculate some information such as its boundary, precise size, localization of tumors, and spatial relations. Final, we reconstructed the extracted lung and tumor regions to a 3D volume module to provide physicians the more reliable vision. In order to validate the proposed system, we have tested our method in a database from 15 lung patients. We also demonstrated the accuracy of the segmentation method using some power statistical protocols. The experiments indicate our method results more satisfied performance in most cases, and can help investigators detect lung lesion for further examination.

A novel idea of automatic recovery from photo QR Code is proposed. The photo QR code can be generated by the proposed pyramid QR coding method. The original photo is then embedded into the generated QR code. The experimental result demonstrates that the original photo is with very high quality and the photo QR code can be used for decoding the embedded words or URLs by common QR code decoders.

Recently, it is very popular to convert image or video to different format. For example, copy a color printed paper to gray scale paper or show video with different devices. This paper proposes a fast removable watermark generating method. The watermark disappears after the watermarked image converts to another format. As a gray image can be generated from different color images with same converting approach, the goal of this paper has been achieved. An example shows that it is possible to create recoverable contour style watermark by control hue and saturation properly. A vision quality optimum approach also presented. The advantages of proposed method includes the recover image is identical to the converted image without watermark, people can remove the watermark without computer or the computational cost is very low. The result of this paper is useful for secret removing or demo video transmitting.

Visible watermarking is a common intellectual property protection mechanism for digital images. It is fulfilled by translucently overlaying a visual logo and/or a segment of text onto an image to declare the ownership of the image and thus deterring unauthorized usage to the image. This paper presents a novel visible watermarking technique by considering HVS characteristics and applying a distance transform function to evaluate appropriate stamping strength for individual watermark pixels. It exhibits lucid watermark pattern under moderate modification to the host image. Compare to uniform-contrast watermarks generated using conventional HVS based visible watermark schemes, the proposed method achieves better balance to the visual quality between the watermark and the image content beneath it.

This paper proposes architecture to hide proverb data in the corresponding animator image. The method not only stores data but also prevents the relation between image and proverb data lose. The hiding data includes the word of proverb, pronunciation and the explanation. Base on the digital data hiding algorithm, the goal has been achieved. A prototype system is also established in this study. The result can help the traditional culture preserve in the internet.

In 2D-to-3D video conversion process, 3D video can be generated from 2D video and its corresponding depth map by depth image based rendering (DIBR). The depth map is the key in the conversion process as it provides immersive experience to viewers. So the copyright protection for depth map must be considered. Traditional unseen visible watermarking (UVW) for depth map protection cannot resist filtering attacks. In this paper, a robust unseen visible watermarking (RUVW) scheme is proposed, in which the watermark regions without interference are detected for embedding, the copyright information is enhanced with Discrete Cosine Transformation (DCT) and watermark can be seen directly when the rendering conditions are changed. The experimental results show that the proposed method has good robustness against various attacks such as scaling, filtering, noises and compression.

The ViBe algorithm is a powerful technique for the background detection and subtraction in video sequences. Compared to the state-of-the-art algorithms, ViBe algorithm is better in fast speed and less memory consumption. However, when applying ViBe algorithm to the moving objects appeared in the first frame of videos, all pixels in first frame will be used to build the background model that will result in the foreground pixels in sample set. This problem causes the ghost areas emerge. And it will remain for a long time. In this paper, a salient region detection based ViBe algorithm is proposed to eliminate the ghost areas fast. First, the foreground region is extracted from the first frame of videos using the salient region detection algorithm. According to the result of salient region detection, the background area of image is separated from the foreground area. The foreground pixels are dislodged from the sample set. Then, only background pixels are used for background model initialization. The experimental result shows that the improved algorithm can eliminate ghost in few frames quickly.

Nowadays, traffic accidents is universally decreasing due to many advanced safety vehicle systems. To prevent the occurrence of a traffic accident, the first function that a safety vehicle system should accomplish is the detection of the objects in traffic situation. This paper presents a popular method called boosted HOG features to detect the pedestrians and vehicles in static images. We compared the differences and similarities of detecting pedestrians and vehicles, then we use boosted HOG features to get an satisfying result. In detecting pedestrians part, Histograms of Oriented Gradients (HOG) feature is applied as the basic feature due to its good performance in various kinds of background. On that basis, we create a new feature with boosting algorithm to obtain more accurate result. In detecting vehicles part, we use the shadow underneath vehicle as the feature, so we can utilize it to detect vehicles in daytime. The shadow is the important feature for vehicles in traffic scenes. The region under vehicle is usually darker than other objects or backgrounds and could be segmented by setting a threshold.

Panoramic image mosaic is a technology to match a series of images which are overlapped with each other. Panoramic image mosaics can be used for different applications. Image mosaic has important values in various applications such as computer vision, remote sensing image processing, medical image analysis and computer graphics. Image mosaics also can be used in moving object detection with a dynamic camera. After getting the panoramic background of the video for detection, we can compare every frame in the video with the panoramic background, and finally detect the moving object. To build the image mosaic, SURF (Speeded Up Robust Feature) algorithm is used in feature detection and OpenCV is used in the programming. Because of special optimization in image fusion, the result becomes stable and smooth.

This paper proposes an evaluation framework for an audio data hiding system that can be used for long-distance aerial transmissions where the stego speech signal is emitted by outdoor loudspeakers of voice evacuation and mass notification systems. Typical disturbances of aerial transmissions are modeled and implemented by signal processing units in series. The results of computer simulations for long-distance (70–800 m) aerial transmissions show that a bilateral time-spread echo hiding combined with a novel frame synchronization technique exhibits better detection performance than the conventional time-spread echo hiding under extremely low signal-to-noise ratios of 0 and −5 dB. Moreover, high-frequency attenuation and frequency shifts at the receiving side degrade the detection performance of the time-spread echo hidings.

Digital watermarking is a technique used for embedding information in digital content and protecting its copyright. The important issues to be considered are robustness, quality and capacity. Our goal is to satisfy these requirements according to the Information Hiding and its Criteria for evaluation (IHC) criteria. In this study, we evaluate our watermarking scheme along the IHC criteria Ver.3 as the primary step. Although image watermarking techniques based on machine learning already exist, their robustness against desynchronization attacks such as cropping, rotation, and scaling is still one of the most challenging issues. We propose a watermarking scheme based on machine learning which also has cropping tolerance. First, the luminance space of the image is decomposed by one level through wavelet transform. Then, a bit of the watermark and the marker for synchronization are embedded or extracted by adjusting or comparing the relation between the embedded coefficients value of the LL space and the output coefficients value of the trained machine learning model. This model can well memorize the relationship between its selected coefficients and the neighboring coefficients. The marker for synchronization is embedded in a latticed format in the LL space. Binarization processing is performed on the watermarked image to find the lattice-shaped marker and synchronize it against cropping. Our experimental results showed that there were no errors in 10HDTV-size areas after the second decompression.

We propose a watermarking method using scale-invariant feature transform (SIFT) features that have both scale and rotation invariance, and evaluate our method in accordance with the information hiding criteria (IHC) ver. 5. It is defined as evaluation criteria against several possible attacks; these attacks are JPEG compression and geometric attacks, e.g., scaling, rotation, and clipping. In our method, we use local feature regions located around the SIFT features that are robust against scaling and rotation. The regions are normalized in size and selected as marked regions. Watermarks are embedded in the marked regions. We also introduce two error-correction techniques: weighted majority voting (WMV) and low-density parity-check (LDPC) code. When a stego-image is attacked by scaling or rotation, the image is spatially distorted. WMV and LDPC code can correct errors of extracted watermarks in the distorted stego-image. On the other hand, it is not easy to detect rotated marked regions. Therefore, the correct orientation is searched for by brute force. We evaluated the proposed method in accordance with IHC ver. 5. Our method can achieve robustness against scaling and rotation attacks in the highest tolerance category.

Among data hiding methods for documents, the spaces between words and paragraphs are popular features for embedding information. However, in order to avoid the visible distortions, the embedding capacity is limited to be small in conventional methods. In this paper, we regard a collection of space lengths as one dimensional feature vector and apply signal processing approaches for the vector to embed more information with less distortions. We also focus on the Portable Document Format (PDF) files and propose a new data hiding method for PDF files with large capacity. Considering the secrecy of the embedded information, a random permutation and dither modulation are introduced in the operation.

In this paper, we propose a novel digital audio watermarking system inspired by the tally trade. In ordinary digital audio watermarking, a single stego signal is produced through the embedding process and the hidden message is extracted from the stego signal. In tally-based system we propose, multiple stego signals are tallied to produce and the hidden message which is extracted from the temporally mixed signal composed of the required number of stego tallies. When the extractor can’t mix the required number of stego tallies, it misses the hidden message. The system’s performances are evaluated in terms of the number of tallies (shares).

We use the adaptive clustering method that is the division of network by following the complex network theory model to manage nodes in case of WSN node scale changes and node data traffic surges, which complicates the node relationship at the same time and puts forward higher requirements for security management. In this paper, it is proposed to generate a variable-length random key to manage the relationship between the nodes to ensure that the complex relationship between the node changes and the security of the entire network, and provide an easy-to-implement variable security management mechanism.

Aiming at improving the real-time performance of infrared weapon systems, a method based on block compressed sensing is proposed to detect infrared small target, which is also easy to be implemented with hardware. The proposed method can detect and locate small infrared targets by classifying compressed results of blocks. In addition, in order to solve the low detection accuracy caused by using uniform block, the proposed method uses overlapping blocks to reduce the maximum distance between the center of the test sample block and that of the target. Experiments show that the proposed method can effectively improve the detection accuracy of infrared small targets.

At present, most of the complex electronic systems use communication protocols for data transmission. In the process of system joint test, the system usually needs to monitor the communication protocol data, that is, protocol data decoding,or replace the equipment in the system in a semi-physical way to debug the system. This requires a semi-physical device to dynamically generate protocol data frame into a complex system, that is, the encoding of the protocol data.

Hyperspectral image Contains rich information, so improving the classification ability of hyperspectral image has become a hot spot of research recently. Image segmentation is one of the most basic and important problems of the image processing and low level computer vision and the precondition of the image processing, some hyperspectral image classification method based on image segmentation. In this study, an image segmentation method based on kernel methods was proposed. First reduce the dimension of hyperspectral images by KPCA, then the k-means method are used to cluster, finally finish the image segmentation in the high-dimensional space by gaussian kernel mapping. In simulation experiment, changing the parameters of the experiment and comparing with standard hyperspectral image segmentation, the results show that the method is good enough in image segmentation, which can be used for visual analysis and pattern recognition, and realize hyperspectral image segmentation.

According to JPEG2000 image encryption algorithm based on confusion of the wavelet coefficients cannot support region-of-interest(ROI) image coding, an encryption method for ROI images is proposed. The proposed method Generates the confusion table using Chaotic sequences, confuses the Wavelet coefficients inside each code blocks, and encrypts the sign bits using Chaotic sequences. The experimental results show that the proposed method has lower affection to the compression ratio, and has the ability to provide adequate key space to resist the brute attack. The proposed method is important to support ROI image coding.