New Approaches for Multidimensional Signal Processing

Recently, computational intelligence (CI) techniques have become efficient intelligent tools for brain tumor detection. It has become one of the major research subjects in medical imaging and diagnostic radiology. In the area of processing the brain images, computer-aided diagnosis (CAD) systems are basically relied on different CI techniques in all its stages to implement a smart consultation system that can help the radiologists by providing a second opinion that can assist in detection and diagnosis of brain tumors. This paper presents a comprehensive and up-to-date research in the area of digital medical imaging covering a wide spectrum of CI methodological and intelligent algorithm. The paper discusses the current research of the CI techniques for developing smart CAD systems. We present two applications for a hybrid intelligent technique for automatic detection of brain tumor through MRI. The technique is based on the following CI methods: the feedback pulse-coupled neural network for image segmentation, the discrete wavelet transform for features extraction, the principal component analysis for reducing the dimensionality of the wavelet coefficients, and the feed-forward back-propagation neural network to classify inputs into normal or abnormal.

Human gait is essential for long-term health monitoring as it reflects physical and neurological aspects of a person’s health status. In this paper, we propose a non-invasive video-based gait analysis system to detect abnormal gait, and record gait and postural parameters framework on a day-to-day basis. It takes videos captured from a single camera mounted on a robot as input. Open Pose, a deep learning-based 2D pose estimator is used to localize skeleton and joints in each frame. Angles of body parts form multivariate time series. Then, we employ time series analysis for normal and abnormal gait classification. Dynamic time warping (DTW)-based support vector machine (SVM)-based classification module is proposed and developed. We classify normal and abnormal gait by characterizing subjects’ gait pattern and measuring deviation from their normal gait. In the experiment, we capture videos of our volunteers showing normal gait as well as simulated abnormal gait to validate the proposed methods. From the gait and postural parameters, we observe a distinction between normal and abnormal gait groups. It shows that by recording and tracking these parameters, we can quantitatively analyze body posture. People can see on the display results of the evaluation after walking through a camera mounted on a companion robot.

In this work are presented two new approaches for hierarchical decomposition represented as tensors of size N × N × N for N = 2ⁿ, based on algorithms which (unlike the famous similar approaches) do not require iterative calculations. Instead, they use repetitive simple calculations in each hierarchical decomposition level. As a result, the computational complexity (CC) of the new hierarchical algorithms is lower than that of the iteration-based. In general, hierarchical decompositions are divided into two basic groups: statistical and deterministic. To the first group is assigned the algorithm hierarchical tensor SVD (HTSVD) based on the multiple calculation of the two-level SVD for the elementary tensor of size 2 × 2 × 2. The decomposition is executed by using the HTSVD in three orthogonal spatial directions simultaneously. The deterministic decompositions have lower CC than the statistical, but they do not ensure full decorrelation between the components of the 3D decompositions. In this group are the famous orthogonal transforms 3D fast Fourier transform (3D-FFT), 3D discrete cosine transform (3D-DCT), 3D discrete wavelet transform (3D-DWT), 3D contourlet discrete transform (3D-CDT), 3D shearlet discrete transform (3D-SDT), etc., and also, the algorithm 3D reduced inverse spectrum pyramid (3D-RISP). The last is distinguished by its lower CC and the high energy concentration in the first decomposition components. To achieve this, for the basic tensor of size 2ⁿ × 2ⁿ × 2ⁿ is executed the 3D fast truncated Walsh–Hadamard transform (3D-FTWHT). Significant advantage of 3D-RISP compared to the famous pyramidal decompositions of the kind 3D-DWT, 3D-CDT, 3D-SDT, etc., is the absence of 3D decimation and 3D interpolation which produce distortions in the restored tensor.

In this paper, we present a comparative analysis of the performance of the Tucker-ALS, CP-ALS, Tucker-ADAL, and the HoRPCA-S tensor decomposition algorithms, applied for tracking of domestic animals in video. Decomposition and full processing time, detection rate, precision, and F-measure are the evaluating parameters revealing the efficiency of each algorithm. Promising results suggest the applicability of the investigated decompositions but also demonstrate particular differences among them in terms of decomposition time and detection rate. In order to increase the detection rate of systems of parallel type employing multiple decomposition algorithms we propose a score fusion with fair voting which performs better than some of the tested algorithms alone.

In real applications, the same object may have been presented by different shapes due to the moment and the angles of image acquisition, which does not guarantee a complete contour extraction without being disturbed by the noise or the distortions. In this paper, we propose a new method to match partially occluded shape based on affine curvature scale space. Firstly, an affine curve re-parameterization is defined, inspired by the properties of affine curvature scale space (ACSS) shape descriptor. Then, the different parts will be matched in order to minimize the \( L_{2} \) distance by the calculation of the pseudo-inverse matrix to estimate the translation and the linear transformation based on the affine curve matching (ACM) algorithm. Finally, a matching curve algorithm is obtained according to any planar affine transformation and in any partial occluded case. Experiments are conducted on multi-view curve dataset.

This paper addresses the problem of vision-based line tracking control of unicycle mobile robots. First, a robot-camera model suitable for path following applications is derived. Using a look-ahead approach, a feedback controller is proposed for tracking curved paths on the ground using information from an onboard down-looking camera using distance-only measurements. Stability properties of the closed-loop system are analyzed, and asymptotic stability of the resulting closed-loop control system is proved using Lyapunov stability theory. Simulation and experimental results are presented to illustrate the effectiveness of the proposed control scheme.

This paper presents the implementation of a markerless mobile augmented reality application called a virtual eye glasses try-on system. The system first detects and tracks human face and eyes. Then, the system overlays the 3D virtual glasses over the face in real time. This system helps the consumer to select any style of glasses available on the virtual space saving both time and effort when shopping online. A method based on local-invariant descriptors is implemented to extract image feature points for eyes detection and tracking. A new approach for camera pose estimation is proposed to augment real images with virtual graphics. Experiments are conducted using Haar cascade and speeded up robust features (SURF) cascade. The system is optimized and adapted for a mobile architecture.

Images can easily be manipulated without any visual marks to the naked human eye with massive improvements in image manipulation software. This tampering is the main propelling force for the need of better image forensics such that field is known as image forgery detection. Any digital image with regions where the image contents are identical is said to have copy–move forgery (CMF). Copy–move forgery is performed to improve the visual features or to cover the underlying truth in the image. Many algorithms have been used for CMF detection, and this work is about improved key-point and clustering-based CMF detection scheme. The proposed scheme combines the efficiency of a key-point-based scheme and clustering of these key points to further improve the results. Modified Harris operator-based key-point detection algorithm with clustering using local gravitation is utilized for key-points selection. The average accuracy, PSNR and SSIM rates are used to evaluate the performance of the proposed algorithm with scale-invariant feature transform (SIFT), which is another state-of-the-art key-point algorithm. The paper concluded with the efficiency of the key-point-based scheme.

The paper presents the development of a Web-based virtual reality (VR) environment for planning and simulation of lifting operations performed by a hydraulic excavator. The excavator is represented as a mechanical system with four degrees of freedom (DOF) consisting of a fixed base body, three-link digging manipulator, and a swinging payload. An inverse kinematic model and a dynamic model of the excavator during performing lifting operations with bucket following prescribed vertical straight-line trajectory are developed. Workspace characteristics of the digging manipulator are determined, and the subset of the workspace corresponding to the cutting-edge positions for constant orientation of the bucket is identified. Using modern Web-based technologies, a VR environment for planning, simulation, and 3D animation of the excavator motion is developed. It consists of an HTML5 Web page with an X3D model of the scene, open-source framework X3DOM, and JavaScript functions. The developed VR environment considerably facilitates the design, investigation, planning, and e-learning of the lifting operations, performed by the excavator.

Colorectal cancer is nowadays the fourth cause of cancer death worldwide. Prevention of colorectal cancer by detection and removal of early stage lesions is of essential importance and has become a public health challenge worldwide. As the screening is carried out mainly by some sort of endoscope, and the endoscopic image processing is an important area of research and development, it is essential to know what kind of measures are used in determining whether polyp finding hit rates or miss rates are acceptable. It is rather natural to match the hit rate measures to the method itself; thus, in this contribution, the most typical polyp detecting methods are summarized shortly together with the metrics they use for evaluation of their results. However, in computer-aided diagnostics, the measure that is used by the medical community might differ from the measures typical in image processing researches. Also, the output of such polyp detecting methods is tested as inputs for active contour methods.

Phishing attack has posed a greater threat to user information over network. In addition to the existence of various disguise illegal URL’s, instances had been seen when users are redirected to phishing URL’s that challenges their privacy concern. In the current work, author tried to develop an effective anti-phishing method based on hybrid similarity approach combining cosine and soft cosine similarity that measures the resemblance between user query and database. The strength of the proposed hybrid approach is further enhanced with the incorporation of feed forward backpropagation neural network (FFBPNN) so as to validate the similarity-based predictions. The model evaluated against 3000 sample files demonstrated to effectively detect phishing attacks with positive predictive value, true positive rate and F-measure of 71.9%, 72.6% and 72.23%, respectively.

The fast improvement of the Web has expanded the simplicity of sharing data to individuals around the world. In any case, this headway additionally raises a trouble about information control when the data is transferred by the sender to the beneficiary. Along these lines, data security is a significant issue in information correspondence. Steganography and cryptography assume significant jobs in the field of data security. Steganography can be applied, however, a different computerized media, for example, video, pictures, and sound to cover data in such a manner that nobody else realizes that there is a secret data. Cryptography alludes to the specialty of changing over a plaintext (message) into an ambiguous organization. Both steganography and cryptography strategies are strong. The contents obtained after doing so is secret and its existence is also hidden. This method is tested and it is observed that it prevents steganalysis too as well as parameters like PSNR and MSE are also tested which gave good results.

In recent years, there is a rise in platforms like YouTube, Amazon, Netflix and many other Web services. Thus, there is a need of recommender systems in our lives. Recommender systems have become unavoidable in online journeys. Also, people are largely depending on online shopping platforms. When people buy from online shopping platforms they search and browse various products for their purchase. And even if they only browse, they can see the products everywhere on the Internet, i.e., on their social media platforms, blogs, etc. So, online shopping platforms like Flipkart and Amazon use recommendation systems to recommend products related to their customers. These platforms collect information about their customer’s pattern of shopping and purchasing behavior to give accurate and needful recommendation to the customers. In this paper, various recommendation systems are discussed. Comparative analysis is also performed for these recommendation systems. It also gives the idea of how these recommendation systems help to improve and manage audience for online shopping platforms. It also discusses that how the similarity can be measured among the items for content-based filtering and similarity among users for collaborative filtering. It also describes that hybrid-based filtering can be used to overcome the disadvantages of content-based and collaborative filtering to give item of interest and other items recommendation to users.

Recently, the smart use of biometric traits, i.e., fingerprints, face, finger knuckle print, etc., in user authentication system seems to be integral part because of their user-friendly and robust behavior. All of these traits present different degrees of uniqueness, permanence, durability, performance, user acceptance, and robustness and are valuable according to their need in respective application. The proposed approach illustrates finger knuckle print biometric (FKP) for the authentication of user as it avoids latent FKP and criminal investigation stigma associated with printing the surface of the knuckles. The proposed approach used public database and the preprocessing carried out on the images collected, in order to separate the index finger, middle, and ring fingers of the hand. Bayesian network is used to extract the feature of FKP for the authentication and identification of user. The image processing is carried out using MATLAB R2014 software.

 Current article introduces a new approach for detection of fires based on deep neural network (DNN) and their extinguishing using an acoustic fire extinguisher. Finding fires on video stream is based on low-cost hardware platform containing Movidius stick for hardware acceleration of the DNN used for fire detection. For this purpose, the fire extinguisher uses a sinusoidal acoustic wave modulated by a triangular waveform. The special design of the extinguisher guarantees that the sound pressure level will be sufficient for successful extinguishing of the fire in distance up to 130 cm.

Under the background that modern information technology and the Internet are closely related to people’s lives, deep learning algorithm based on convolutional neural network has important application significance in image processing. At present, the development of the top science and technology in various industries in society is often related to artificial intelligence, convolutional neural network algorithm applied to electronic information image processing provides a means and approach for artificial intelligence, and it has been widely used for image analysis in the medical field. The fields of intelligent recognition in computer systems and even image restoration in criminal investigation have contributed to the development of high-quality scientific and technological life in modern society. Based on the above background, this article gives a brief introduction to the related technologies of convolutional neural network applications and electronic information image processing. I hope that those who are interested in research in the related fields will have some knowledge.

Pharmaceutical logistics distribution is a kind of physical transportation method that sends drugs from manufacturers to relevant medical units and departments. It is based on the extensive application of traditional logistics systems in the field of medicine. Since logistics distribution is a way to directly contact consumers, it largely determines the service level and operating costs of medical companies. Therefore, logistics distribution is also an important part of the logistics link, and it has been widely valued and studied by relevant parties. In addition, to show a good service attitude in the logistics system how to reduce operating costs is also a key issue. Therefore, the particle swarm intelligence optimization algorithm is used to reasonably plan the distribution path of transportation vehicles in logistics transportation to effectively improve transportation efficiency. Reducing transportation costs to a greater extent has extremely important theoretical significance and use value for the research and application of logistics.

Among researches on parameter extraction in sheep body measurements, wool is one of the important influence factors causing sheep body measurement errors. Wool images were analyzed in this study using ridge regression algorithm, KNN algorithm, and SVM to discriminate wool length. The final body measurement parameters were processed according to the discrimination results in order to reduce the measurement errors caused by non-uniform wool length.

The dynamic diffraction patterns caused by an asymmetric liquid-core cylindrical lens filled with diffusion solution are observed using CCD and simulated using MATLAB based on Collins formula in this paper. The diffusion coefficient of ethylene glycol diffusing in pure water is measured at 298.15 K by analyzing the movement of the narrowest position in diffraction patterns base on Fick’s second law. The correctness and application conditions of this measurement method are further interpreted by comparing the simulation and experimental diffraction patterns for the first time. The work expressed in this paper deepened the understanding of the scalar diffraction theory, diffusion process and Fick’s second law, which paved a way for the further research on diffusion phenomenon.

A new approach for multidimensional graphic objects filtration using HoSVD tensor decomposition is presented. The experimental studies were performed on a set of test 3D images of size of 100 × 100 × 100, containing simple geometric objects—a sphere, a cylinder, a cone, etc. After that, Gaussian noise with different variation is added and the low-frequency part of the decomposition matrices U and S is filtered. The results obtained of the filtered images show that the quality of the restored images in the different planes and in total for the 3D image is excellent. The peak signal-to-noise ratio for different samples ranged from 28 to 50 dB.