Search

This paper presents a digital image encryption scheme based on multi chaotic system and random DNA coding. Firstly, the initial values and parameter values of 2D Logistic-adjusted-Sine mapping (2D-LASM) and Logistic-Sine system (LSS) are obtained from SHA256 hash values of the original image. In the scrambling stage, the chaotic sequences generated by 2D-LASM are used to get two column scrambling matrices and row scrambling matrix, respectively. The elements of the second column scrambling matrix and row scrambling matrix are used as row and column coordinates in the scrambling process. Then apply it to scrambling the DNA encoded image, which can complete the row and column scrambling at the same time. In the diffusion stage, we proposed a new diffusion method. Through the chaotic sequence generated by the LSS, a DNA coding sequence is obtained. DNA XOR operation is carried out on the central point and horizontal line of the image, and then spread from the central line to the upper and lower directions of the matrix to achieve the purpose of multi-directional diffusion and improve the encryption efficiency. Experimental results and security analysis show that the algorithm executes fast and has strong security. It can resist many attacks, such as statistical attacks, brute attacks, plaintext / select plaintext attacks, etc. [ABSTRACT FROM AUTHOR]

Paper checks may have complex background features including fine lines and patterns, which make forgery more difficult. Also, halftoning techniques are used to produce continuous tones and to prevent copies with void pantograph features. When these kinds of checks are scanned, Moire patterns may occur. These patterns make it difficult for customers to examine the scanned check images on ATM (Automated Teller Machine) displays. They also can decrease the classification accuracy of check recognition systems. In this paper, we propose an algorithm to enhance the perceptual quality of scanned check images by reducing the Moire patterns. The proposed algorithm consists of foreground extraction, Moire detection and Moire removal. Subjective image quality assessment was performed to evaluate the degree of improvement. Experimental results show that the proposed algorithm improves perceptual quality while maintaining check recognition accuracy. [ABSTRACT FROM AUTHOR]

The purpose of the digital image self-recovery is to restore high quality images as much as possible when the image is tampered. Existing algorithms can only achieve high recovered image quality with tiny tampering rate. To obtain high recovered image quality with large tampering rate, this paper proposes a digital image self-recovery algorithm based on improved JSCC (Joint Source-Channel Coding) optimizer. The algorithm performs quadtree decomposition of original grayscale image corresponding to different decomposition factors γ and performs bit-plane layering according to the block class obtained by quadtree decomposition. Size of each bit-plane is the number of pixels of each block class. Then, the original image is compressed by SPIHT, and the compressed bit-stream of SPIHT is segmented into bit-plane according to the size in order. The bit-planes are protected by different RS (Reed-Solomon) coders to get optimal decomposition result of corresponding γ. Finally, JSCC optimization is designed to get an optimal quality of recovered image. Experimental results show that, using our algorithm, for 2-LSB embedding, when the tampering rate is less the minimum TTR (Tolerable Tempering Rate), the PSNR is improved by 4.93dB. When the tampering rate is larger the minimum TTR, the PSNR is improved by 2dB. When 3-LSB watermark are embedded, the PSNR of recovered image is improved by 2.58dB on average. It shows that our improved optimizer effectively improves the quality of the recovered image at high tampering rates, compared with the similar algorithms. [ABSTRACT FROM AUTHOR]

Fog computing is considered a formidable next-generation complement to cloud computing. Nowadays, in light of the dramatic rise in the number of IoT devices, several problems have been raised in cloud architectures. By introducing fog computing as a mediate layer between the user devices and the cloud, one can extend cloud computing's processing and storage capability. Offloading can be utilized as a mechanism that transfers computations, data, and energy consumption from the resource-limited user devices to resource-rich fog/cloud layers to achieve an optimal experience in the quality of applications and improve the system performance. This paper provides a systematic and comprehensive study to evaluate fog offloading mechanisms' current and recent works. Each selected paper's pros and cons are explored and analyzed to state and address the present potentialities and issues of offloading mechanisms in a fog environment efficiently. We classify offloading mechanisms in a fog system into four groups, including computation-based, energy-based, storage-based, and hybrid approaches. Furthermore, this paper explores offloading metrics, applied algorithms, and evaluation methods related to the chosen offloading mechanisms in fog systems. Additionally, the open challenges and future trends derived from the reviewed studies are discussed. [ABSTRACT FROM AUTHOR]

In order to reduce the impact of disguised missing value on data analysis, this paper proposes a new algorithm - The Detection Algorithm for Disguised Missing Value Based on Filter-Kmeans. The algorithm identifies mainly disguised missing value by the clustering effect, and is applied to the data set with certain probability of data points. Firstly, the suitable data object points are selected using the silhouette coefficient method and Bernoulli's law of large numbers. And then the weighted average distance is used to control the cluster traversal. Finally, the filtering operation is performed during the process of cluster traversal. According to the experimental results, the algorithm achieves better improvement in the precision ratio, recall ratio and F1-measure on the open dataset. [ABSTRACT FROM AUTHOR]

S-box is the only non-linear device in the cryptographic algorithm, and its quality determines the lower limit strength of the cryptographic algorithm. However, because the image data is highly correlated, the traditional encryption methods and their S-boxes, such as AES and DES are not suitable for using in image encryption. Based on this, this paper proposes an S-box generation algorithm based on a 4D hyperchaotic system and improved particle swarm optimization. Firstly, this paper improves on the Lorenz chaotic system and proposes a 4D hyperchaotic system with a higher Lyapunov exponent and more complex dynamics. Secondly, the idea of simulated annealing algorithm is introduced into the particle swarm optimization algorithm, which further improves the efficiency of the particle swarm optimization algorithm and improves the problem that the particle swarm optimization algorithm is easy to fall into the local optimal solution. Then an improved particle swarm optimization algorithm is used to optimize the nonlinearity of the S-box to improve the performance of the S-box. Finally, use the generated S-box to design an image encryption algorithm and prove the security of the S-box. The experimental results show that the S-box designed in this paper has excellent performance in the five indicators of Nonlinearity, SAC, BIC-NL, LP, and DP. At the same time, the encryption result can resist common attacks, so it has strong multimedia security. [ABSTRACT FROM AUTHOR]

This paper proposes a new chaotic-based encryption algorithm to encrypt images securely and effectively. A combination of Arnold cat, logistic mapping and image blocking has been used to design this algorithm. The proposed algorithm is a special algorithm that not only does it use image blocking once to relocate the pixels in four areas of the image separately but also uses an Arnold cat mapping to relocate the pixels across the whole image once again. This results in a higher number of pixel relocations which makes it impossible to decode without the key. The other interesting and new point is the particular use of chaotic logistic mapping to develop three image keys and to combine them with the input image, and to develop four other image keys to encrypt the image. Relocation of the input image blocks and the images generated by the logistic mapping are conducted through sorting the numbers generated by the mapping and how they have been relocated. Then, Arnold cat is used to relocate the pixels across the whole image (which is a combination of the input image and three images generated by logistic mapping). Eventually, the encryption operation is conducted on the obtained image using the encryption key (the image generated through the combination of the four keys). Numerous statistical tests and security analyses indicate the excellent security of our proposed algorithm. [ABSTRACT FROM AUTHOR]

The recognition of abnormal behavior in surveillance video is the focus of current research, which has high research value and broad application possibilities. Its main applications are in the fields of intelligent surveillance, intelligent security, and smart cities, and it is of great significance to study the recognition of abnormal behaviors. Because of the complexity of human movement and the variability of the external environment, the recognition and detection of abnormal behaviors have some challenges. The recognition and detection of abnormal human behaviors in surveillance video still needs further research and development. This paper uses the multi-branch convolutional neural network to extract the spatial features of video frames for the first time, and as an encoder to pass the condensed features to the Gated Recurrent Unit (GRU), which extracts Temporal features from multiple video frames. And then the Gated Recurrent Unit output the result as the decoder. We did a series of comparative experiments on UCF-Crime dataset. And finally, we achieved an accuracy of 86.78% in the test set. The experimental results show that our multi-branch convolutional fusion neural network is better than previous surveillance video abnormal behavior recognition algorithms. At the same time, in order to verify the generalization performance and efficiency of the algorithm, we also conducted an experimental validation on the UCF-101 dataset in this paper, and the results show that the algorithm in this paper can also show a high accuracy rate on the UCF-101 dataset, and the speed of the algorithm is almost close to that of the C3D method with improved accuracy rate, making it possible to develop simple recognition applications based on the algorithm studied in this paper subsequently. [ABSTRACT FROM AUTHOR]

In recent years, two similar formulations have been proposed for semi-blind robust SVD-based watermarking. Several research articles reported some faults in those formulations, which are known collectively as the False-Positive Problem. Other researchers investigated different aspects of this problem from different viewpoints, and proposed solutions to overcome the faults. This paper, first categorizes those researchers' viewpoints into three groups, and shows that while focusing on certain aspects of the two algorithms, they did not succeed in finding the real cause of the False-Positive Problem. Secondly, this paper locates the actual cause of the faults of the two algorithms by directly examining the formulations. Our experiments show that the algorithms fail to correctly embed a distinguishable part of the watermark into the cover image. Further analysis shows that the algorithms embed some arbitrary data in the cover image, and send most of the watermark data as the side information in a semi-blind watermarking setup. Finally, this paper reviews the existing solutions to the False-Positive Problem, and shows that the correct solution to this problem is embedding the principal component of the watermark into the cover image. [ABSTRACT FROM AUTHOR]

Background subtraction is a prerequisite for a wide range of applications, including video surveillance systems. A significant number of algorithms are often developed and published in different publication mediums in the area, such as workshops, symposiums, conferences, and journals. An important task in presenting a new background subtraction algorithms is to clearly show that its performance outperforms the performance of the state-of-the-art algorithms. In this paper, we present recommendations on how to evaluate the performance of background subtraction algorithms for surveillance systems. We identified, through a systematic mapping, the key steps and components of this evaluation process – procedures, methods, and tools – most used by the authors in each of these steps. Considering this statistical analysis, we perform a theoretical analysis of the most used key components to identify their pros and cons. Then, we define a set of recommendations that aim to standardize and clarify the performance evaluation process of a new background subtraction algorithm. [ABSTRACT FROM AUTHOR]

Data clustering is one of the branches of unsupervised learning and it is a process whereby the samples are divided into categories whose members are similar to each other. The K-means algorithm is a simple and fast clustering technique, but it has many initial problems, for example, it depends heavily on the initial value for better clustering. Moreover, it is susceptible to outliers and unbalanced clusters. The artificial bee colony (ABC) algorithm is one of the meta-heuristic algorithms that is used nowadays to solve many optimization problems including clustering and the fundamental problem of this algorithm is exploration and late convergence. In this paper, to solve the problem of exploration and late convergence in ABC are used Random Memory (RM) and Elite Memory (EM) called ABCWOA algorithm. RM in the ABCWOA algorithm has used the search stage for the bait in the whale optimization algorithm (WOA) and EM is also used to increase convergence. In addition, we control the use of EM dynamically. Finally, the proposed method was implemented on ten standard datasets from the UCI Machine Learning Database for evaluation. Moreover, it was compared in terms of statistical criteria and analysis of variance (ANOVA) test with basic ABC and WOA, vortex search (VS) algorithm, butterfly optimization algorithm (BOA), crow search (CS) algorithm, and cuckoo search algorithm (CSA). The simulation results showed that the degree of convergence maintained its performance by increasing the number of repetitions of the proposed method, but the ABC algorithm has shown poor performance by increasing the repetition of performance. ANOVA results also confirmed that the ABCWOA algorithm has a positive effect on the population and it contains less noise than other comparative algorithms. The ABCWOA algorithm show that the ABCWOA algorithm performs better than other meta-heuristic algorithms. [ABSTRACT FROM AUTHOR]

Underwater environment is complex and changeable. In order to obtain more underwater environment information. The larger the field of view of underwater images collected by ROV, the more information is contained. Effective methods to obtain large field of view include fish-eye lens and image stitching. In order to obtain larger field information, we combine the two methods and propose a splicing algorithm that can be applied to fish-eye lenses. This algorithm includes two parts, the first part is correction the fish-eye images, on the basis of the traditional chessboard correction method to improve, this paper put forward a new adaptive gray level method, this method can keep more angular point features, can be more accurate extraction of checkerboard angular point, will be further accurate correction result. In order to achieve real-time underwater patchwork effect. For stitching the corrected images, this paper proposes a fast stitching algorithm (FASTITCH), in the process of image stitching, the algorithm can preserve image feature points and transposed matrix of image matching, so as to calculate the new coordinates, joining together the original feature points in the image. Using this coordinate to match the feature points of another image can save the time of finding feature points in the stitching image, and finally speed up the stitching and complete the task of real-time stitching. The experiment proves that: The error obtained by the new correction method is smaller. Compared with the traditional feature point stitching algorithm, the fast stitching algorithm (FASTITCN) proposed in this paper can shorten the stitching time by about 20%. [ABSTRACT FROM AUTHOR]

In the field of digital imaging systems, autofocus plays increasingly a vital role as a key technology. Autofocus poses a great challenge due to nosiy background and slow focusing speed. This paper presents a new focusing algorithm based on improved Laplacian operator and mountain-climb search algorithm. The clear image after focusing is more different in gray scale than the image without focusing, an image definition evaluation function combining local variance and Laplacian operator is proposed. Learning from the advantages of two-stage recognition in deep learning image recognition, an two-stage search algorithm based on mountain-climb search is designed to better fit the focusing curve near the extreme value of focusing evaluation function, improved mountain-climb search algorithm is divided into rough focusing and fine focusing. The method of rough focusing is used to determine a small focus area, and then fine focusing based on function approximation can greatly improve the efficiency of focus position.The experimental results indicate that this algorithm in this paper is superior to the traditional algorithm in time and accuracy, and the time of the autofocus is reduced by 76%. [ABSTRACT FROM AUTHOR]

In order to improve the management efficiency of software vulnerability classification, reduce the risk of system being attacked and destroyed, and save the cost for vulnerability repair, this paper proposes an automatic algorithm for Software Vulnerability Classification based on convolutional neural network (CNN) and gate recurrent unit neural network (GRU), called SVC-CG. It has conducted a fusion between the models of CNN and GRU according to their advantages (CNN is good at extracting local vector features of vulnerability text and GRU is good at extracting global features related to the context of vulnerability text). The merger of the features extracted by the complementary models can represent the semantic and grammatical information more accurately. Firstly, the Skip-gram language model based on Word2Vec is used to train and generate the word vector, and the words in each vulnerability text are mapped into the space with limited dimensions to represent the semantic information. Then the CNN is used to extract the local features of the text vector, and the GRU is used to extract the global features related to the text context. We combine two complementary models to construct a SVC-CG neural network algorithm, which can represent semantic and grammatical information more accurately to realize automatic classification of vulnerabilities. The experiment uses the vulnerability data from the national vulnerability database (NVD) to train and evaluate the SVC-CG algorithm. Through experimental comparison and analysis, the SVC-CG algorithm proposed in this paper has a good performance on Macro recall rate, Macro precision rate and Macro F1-score. [ABSTRACT FROM AUTHOR]

Elliptical curves are dedicated for several security applications including Radio Frequency Identification (RFID) devices, smart cards, bankcards, etc. To guarantee effective security of such applications, these cryptographic systems require effective resistance to various types of physical attack. Differential Power-Analysis (DPA) attacks were considered the most efficient attacks against scalar multiplication calculation algorithms. In this paper, we propose a countermeasure method against the DPA attacks, for a scalar multiplication algorithm that is basically secure against Simple Power Analysis (SPA) and safe-error attacks. Our proposal is intended for Elliptic Curves Cryptosystems (ECC) algorithms dedicated to low cost applications. We first introduce the different types of side-channel attacks that ECC-based cryptographic algorithms can suffer, as well as their countermeasure methods existing in the literature. We then present an optimized hardware implementation of the most effective scalar multiplication algorithm against SPA and safe-error attacks. Finally, we present our proposed DPA countermeasure method and its effectiveness against other extensions of DPA attacks. Our proposed method is similar to the Basic Random Initial Point (BRIP) method except that the latter is only applicable for the left-to-right algorithm. The proposed method is based on the randomization of processed data during the computation of the scalar multiplication algorithm and prevents vulnerability to Zero-value Point Attack (ZPA), Refined Power analysis (RPA) attack and double attack. In the last part of our paper, we present comparative analysis in terms of computational cost between our proposed method and other countermeasure algorithms presented in the literature, such as Montgomery-ladder, the BRIP algorithm, the left-to-right algorithm and the Co-Z Mont-Ladder algorithm. [ABSTRACT FROM AUTHOR]

In this paper, a new meaningful ciphertext compression encryption scheme is proposed. Firstly, this paper proposes an adaptive sparsification compressed sensing (CS) to get a better reconstructed image by adaptively assigning a more appropriate sparsity threshold to each image. Secondly, considering the noise-like distribution of the measured values, this paper proposes a histogram-shifted least significant bit (LSB) Xor embedding method to reduce the variation of the cover image. The algorithm can also adaptively find the optimal histogram shift parameters based on the distribution of measured values to optimize the ciphertext quality. The peak signal-to-noise ratio of the ciphertext reaches more than 48dB, which is much higher than the existing optimal algorithm. Then, we also proposed a new two-dimensional infinite collapse hyperchaos (2D-IICM), whose two Lyapunov exponents (LE) both reach more than 4 and are in a hyperchaotic state. The sample entropy and information entropy of the sequence can also reach more than 2 and 7.9 respectively, which is 0.1 and 0.3 points higher than other chaos. Good chaos properties provide higher security to the algorithm. This paper uses SHA256 of plaintext to generate a dynamic key, so that the whole encryption system achieves the encryption effect of one-image and one-key, making the entire system highly resistant to attacks. Codes are available at https://github.com/jsw1995/ICMCE [ABSTRACT FROM AUTHOR]

Considering the applications of requiring high fidelity of multimedia content, and for the purpose of copyright protection, a separable and dual data hiding algorithm is proposed. In this paper, the definitions of Ternary Segmentation (TS), Hidden Pixel Pair (HPP), ZN-shape space-filling curve, and Lucas-Arnold Scrambling (LAS) are presented for the first time. By virtue of without degrading the quality of the host image, this paper combines reversible data hiding and zero data hiding to solves the contradiction between robustness and imperceptibility. Firstly, 1-level reversible data hiding is performed using TS, HPP and Huffman compression. Then, 2-level zero data hiding is adopted based on LAS, IWT (Integer Wavelet Transform), BN-SVD(Boost Normed Singular Value Decomposition) and ZN-shape space-filling curve. Finally, separable decryption and extraction is conducted. The experiment results prove that the proposed scheme achieves a better image perceptual quality with an average PSNR of about 49 dB, and stronger robustness with an average NCC of about 0.99 under various attacks. Moreover, it achieves higher security, and the average values of information entropy, correlation coefficient, NPCR and UACI are 7.9, 0.01, 99.6854% and 33.4378%, respectively. [ABSTRACT FROM AUTHOR]

This paper proposes a fast and unified encryption and decryption algorithm based on a composite chaotic system. By combining Logistic map and Sine map, the New-Logistic-Sine map (NLS map) is obtained. NLS map generated the diffusion key matrix needed in the algorithm process, which can enhance the anti-attack ability of the encryption algorithm. Different from most image cryptography systems, the algorithm adopted in this paper has the same encryption process and decryption process, which can save half of the resources in real applications. Firstly, the Secure Hash Algorithm 256 (SHA256) value of the original image was obtained, and the initial values and control parameters of NLS map and Logistic map were calculated; Secondly, the diffusion key matrix is obtained by iterative the NLS map, and is used to perform the first diffusion of the original image; Thirdly, the permutation key sequence is obtained by iterative the Logistic map, and using the sequence to perform the permutation operation on the image after the first diffusion; Finally, the same diffusion key matrix as the first diffusion operation is used to carry out the second diffusion operation on the displaced image to obtain the final encrypted image. The simulation experiment and security analysis show that the proposed image cryptosystem possessed identical encryption process and decryption process, and the algorithm speed is improved ensure the security of the algorithm. [ABSTRACT FROM AUTHOR]

Today according to social media, the internet, Etc. Data is rapidly produced and occupies a large space in systems that have resulted in enormous data warehouses; the progress in information technology has significantly increased the speed and ease of data flow.text mining is one of the most important methods for extracting a useful model through extracting and adapting knowledge from data sets. However, many studies have been conducted based on the usage of deep learning for text processing and text mining issues.The idea and method of text mining are one of the fields that seek to extract useful information from unstructured textual data that is used very today. Deep learning and machine learning techniques in classification and text mining and their type are discussed in this paper as well. Neural networks of various kinds, namely, ANN, RNN, CNN, and LSTM, are the subject of study to select the best technique. In this study, we conducted a Systematic Literature Review to extract and associate the algorithms and features that have been used in this area. Based on our search criteria, we retrieved 130 relevant studies from electronic databases between 1997 and 2021; we have selected 43 studies for further analysis using inclusion and exclusion criteria in Section 3.2. According to this study, hybrid LSTM is the most widely used deep learning algorithm in these studies, and SVM in machine learning method high accuracy in result shown. [ABSTRACT FROM AUTHOR]

Aiming at the problems of image quality degradation and information loss in images affected by reflections in real scenes, this paper proposes a fast and effective specular reflection image enhancement algorithm. This method uses the dark channel prior algorithm to process the specular image, in which the moving window minimum filter is used to estimate the global illumination component of the specular image, and a weighting function based on local pixel chromatic aberration is introduced under the boundary constraints. Then, it uses an improved guided image filtering algorithm to enhance the image, introduces adjustment parameters based on local variance information in the cost function of the guided filtering algorithm, and introduces an adaptive magnification factor in the detail layer. Finally, we compare the algorithm in this paper with the existing algorithms in subjective vision and objective evaluation. The results show that the calculation speed of this method is faster, and it can effectively enhance the information in the reflection image, and simultaneously effectively improve the clarity of the image. [ABSTRACT FROM AUTHOR]

Visual sensors are widely deployed in many promising detection fields, such as precision measurement, visual object detection and search. Instance search is a special object search method based on visual sensor, which is of fundamental importance to many applications, including the area of civil and military. Although a quantity of instance search solutions has been proposed, it remains a critical challenge to provide a real-time, lightweight labeling, easy to deploy and able to simultaneous-object-positioning-and-classification method. To address the challenge, a novel instance search method named Faster-FIIS-GMS is described in this paper. The improved Faster R-CNN with the MRFR (Multi-Receptive Field Residual module) and the APM (Adaptive Proposals Mechanism) is proposed to achieve more accurate and time-saving object detection and background information elimination simultaneously. Next, a novel image matching method named Fast Intensity Initial Screening Algorithm (FIIS) is introduced to filter foreground information in an image preliminarily and then combined with Grid-based Motion Statistics (GMS) to achieve the final searching. The performance of Faster-FIIS-GMS is analyzed and validated by experimental tests in a dataset annotated by manual, and the experimental results demonstrate that the proposed method has great potential in instance search for real scene environment, as it achieves high accuracy and satisfactory real time performance using the general computer. We also provide the comparison to state-of-the-arts to investigate the superiority of the proposed scheme at the end of the work. [ABSTRACT FROM AUTHOR]

Context: The recent research trend has highlighted that multiple stakeholders are involved during requirement gathering in agile software development. Hence, leading to an increased number of duplicate user stories in agile product backlog during requirement gathering. Objective: The objective of this paper is to evaluate the existing techniques employed in identifying and eliminating the duplicate user stories from agile product backlog and to overcome the existing gaps with the help of a newly proposed clustering algorithm. Method: An agile user story is expressed as a function of input and output parameters. That said multiple user stories having similar set of input parameters are most likely to be duplicate causing a redundancy. The newly proposed algorithm is used for clustering user stories having similar set of input parameters through various iterations and then removing the identified duplicate user stories from agile product backlog. This paper also introduces the concept of mass clustering which means clustering a number of user stories in single run. Results: Experimental results prove the proposed model is capable of handling small and large releases ranging between 100 to 1000 user stories with similar efficiency. The proposed clustering algorithm outperformed the clustering algorithms and resulted in 37% decrease in agile product backlog by eliminating duplicate user stories causing redundancy. The experimental results are obtained from the logs of the MATLAB tool. However, the provided algorithm is generic in nature and can be implemented using R, Python or SAS programming tools. The provided algorithms employs proven matrix operations. Conclusion: The proposed clustering algorithm overcomes the limitation of existing user story management methods and clearly out performs when compared with other clustering algorithms. Finally, this paper gives recommendations about the usage of the provided clustering algorithm during agile release planning for eliminating duplicate user stories from agile product backlog. [ABSTRACT FROM AUTHOR]

The workpiece positioning and defect segmentation are two key steps in the workpiece detection process. This paper has designed a CPU-GPU parallel call algorithm based on real industrial quality inspection conditions to realize the high-speed workpiece defect detection. The algorithm can fully utilize all computing resources of the industrial control computer while simultaneously accomplishing the workpiece positioning and defect segmentation tasks. Moreover, to reduce the workpiece defect detection's scope and improve the defect segmentation algorithm's efficiency, the proposed method uses the workpiece positioning results. As for the positioning task, we have designed the double pyramid method to enhance the positioning speed. When it comes to the defect segmentation task, we have introduced the lightweight network to improve the workpiece segmentation speed. Considering that the current general data sets are of the workpiece local image(s) post cutting, we set up a new data set to reflect the situation in the industrial field. It consists of images taken from real industrial fields that can better verify the whole quality inspection algorithm process, including the positioning and segmentation algorithms. According to our experiment, our algorithm accomplished the positioning and defect segmentation tasks at a speed of 116FPS. Additionally, the segmentation accuracy reached 75.12% Mean IoU. [ABSTRACT FROM AUTHOR]

Theoretically, the Super-Resolution (SR) reconstruction scheme is a method which is performed by many applications nowadays for the purpose of generating a High-Resolution (HR) image using the input Low-Resolution (LR) images by filling in the missing high frequency information. In addition, the SR reconstruction implemented based on the theory of sparse representation techniques is known as an effective way to produce HR images using images patches generated from the LR images. In order to improve the quality of denoised images produced by using the sparse representation techniques, a scheme called dictionary learning algorithms could be considered. Thus, the objective of this paper is to provide a performance comparison on the effectiveness of applying the dictionary learning steps with sparse representation algorithms in producing a better denoised image. In this case, the average Peak Signal-to-Noise ratio (PSNR) and Structural Similarity Index Metric (SSIM) values of the denoised image obtained by using Algorithms 1, 2, and 3 which combined the use of dictionary learning and sparse representation algorithms were compared with the values obtained from images produced by applying only sparse regularisation methods. As a conclusion, the denoised images produced by Algorithm 1 in this paper had the greatest average PSNR and SSIM values. Hence, the algorithm with the implementation of the dictionary learning process with sparse representation methods is able to achieve a better result in enhancing the low-resolution images. [ABSTRACT FROM AUTHOR]

The crowding in bus is an important factor affecting passenger satisfaction and bus dispatching level. However, how to use video images to detect crowding accurately is a difficult problem. In this paper, firstly, an image sample library is established based on the evaluation standard of crowding in bus, which contains 16346 sample images. Then, Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM) are used to extract the texture features of the image in bus. Then, a rough classification method of crowding based on Support Vector Machine (SVM) is proposed. At the same time, in order to improve the accuracy of rough classification of crowding, the optimization effects of grid search algorithm, particle swarm optimization algorithm and genetic algorithm on SVM parameters are compared. The results show that the optimization effect of genetic algorithm is the best, and the accuracy rate is 93.20%. Finally, for the problem that the SVM method is not ideal in the fine classification of crowding, this paper proposes a new method based on ResNet. SGD, Adadelta and Adam are selected to optimize the parameters of ResNet model. The accuracy of the optimal Adam algorithm reaches 96.22%, which effectively solves the problem of the fine classification of crowding in bus. [ABSTRACT FROM AUTHOR]

Addressing the challenges of high model complexity, low generalization capability, and suboptimal detection performance in most algorithms for crop leaf disease detection, the paper propose a lightweight enhanced YOLOv8 algorithm. First, by incorporating the advantages of GhostNet's feature redundancy reduction and MobileNet's ability to fuse diverse channel features using the concept of Group Convolution, the paper enhance the backbone network. This improves the network's ability to extract critical features from a multitude of similar redundant feature maps. Second, to improve detection accuracy while reducing model parameters and computational load, the paper introduce the slim-Neck module. Finally, addressing the issue where detection head parameters and computations account for over half of the model, the paper restructure the Head using the concept of shared parameters and integrate convolution blocks that enhance multi-scale information recognition. Results from multiple experiments, averaged for consistency, indicate that compared to the original YOLOv8 algorithm, the improved algorithm achieves an increase in mAP50 from 86% to 87.3% and mAP50:95 from 67% to 68.6%. The model's size is a mere 5.45 MB, and the computational parameter GFLOPs is reduced from 8.1 to 5.5, even lower than the most lightweight YOLOv5. In comparison to other large-model algorithms, this model also demonstrates strong competitiveness in detection accuracy. [ABSTRACT FROM AUTHOR]

In this paper, a new two-dimensional discrete hyperchaotic system is proposed. Verification using the phase-track map, bifurcation map and Lyapunov exponent of the chaotic system shows that the chaotic system has better traversal, hyperchaos, unpredictability and wider chaotic region compared with the existing two-dimensional discrete chaotic systems. To improve the efficiency and security of the image encryption algorithm, this paper firstly uses parallel compressive sensing-aware image processing to greatly improve the efficiency of the image encryption. Secondly, index scrambling and forward and backward diffusion using a combination of a subset of discrete data sequences generated by a two-dimensional discrete hyperchaotic system improves the security of the image encryption algorithm. In addition ,the initial key and the sum of some pixel values of the original image are used to generate the initial values and control parameters of the 2D-SLS chaotic sequence using the hash of SHA-512, making the proposed algorithm robust to known plaintext and selected plaintext attacks. Simulation and experimental results show that the proposed algorithm is more efficient and secure than recently proposed encryption algorithms. [ABSTRACT FROM AUTHOR]

In this paper, the generation of new dragon fractal shapes with chaotic iteration parameters is introduced as the main component of a new efficient approach for different cryptographic applications. This process involves applying a chaotic map, which is considered the initiator pattern, to generate different chaotic dragon fractal (ChDrFr) shapes in lieu of lines (which are classically used to generate dragon fractals). This is the new concept of this paper. The used chaotic maps are sensitive to their initial conditions and are characterized by randomness; hence, the resulting scheme is highly secure. As the resulting ChDrFr shapes have sparse structures, the spaces are packed with random values generated from another 5D hyper chaotic map. For encryption applications based on the substitution approach, one of the five generated ChFrDr shapes can be used to construct a chaotic fractal (ChFr) S-Box, while the other four ChDrFr shapes can be used for diffusion purposes. As an application to these new ChDrFr shapes and the ChFr S-Box, we introduce in this paper a new highly secure image encryption algorithm. A Henon chaotic map is used as the initiator of the ChDrFr shapes. The integer wavelet transform (IWT) is used to generate an approximation and three detail sub-bands for the original image. As the approximation sub-band contains a considerable amount of information about the original image, the above-described ChFr S-Box is used as a replacement for each pixel's value in this sub-band. Then, the resultant substituted image is diffused with one of the generated ChFrDr shapes. The other three ChDrFr shapes are XORed with the details sub-images. Numerical simulation is applied to ensure the efficacy of encrypted images against different attacks. In particular, the correlation coefficient between the initial and the generated images is shown to be nearly zero. Moreover, tests reveal that the information entropy of the encrypted images and UACI were close to their optimum values. The properties of the newly proposed ChDrFr-based encryption algorithm are compared to the ones obtained by other encryption algorithms, and the results prove the superiority of this newly proposed algorithm to other types of encryption methods. [ABSTRACT FROM AUTHOR]

Multi-secret image sharing techniques (MSS) extend the secret image sharing schemes. The paper reports a novel MSS scheme designed to be collusion resistant. It starts with a brief introduction of the concept thereafter discussing the related works in the domain. Security is an implicit requirement for MSS techniques. There should be no possibility of collusion among the shareholders. No less than k shareholders for (k, n) MSS scheme should be having even a slight view of the secret from their shares, n being the total number of shareholders. The paper reviews and highlights the security issues like possibility of collusion, randomness issues or robustness against tampering attack in the state-of-the-art works. It proposes a new (n, n) MSS scheme in order to rectify the issues. The algorithm encrypts N different secret images to generate N noisy share images. It resolves the existing problems in a secure, and time efficient manner. A range of experiments specifically time complexity, noise, tampering, collusion and histogram analysis are performed for rigorous security analysis. The results, thus obtained indicated that the algorithm is efficient and preserves the security requirements in a much better way than the existing MSS schemes. [ABSTRACT FROM AUTHOR]

The new generation versatile video coding (VVC) standard brings extremely high compression efficiency. Meanwhile, the complexity of the encoder is also greatly increased. In view of this, this paper proposes a fast algorithm for all zero block detection to speed up the quantization process, thereby reducing the complexity of the encoder. The method proposed in this paper consists of three parts. Firstly, genuine all zero blocks are detected based on a fixed threshold which is derived by hard decision quantization formula. Secondly, parts of pseudo all zero blocks are detected based on an adaptive threshold which is derived by analyzing the positions of transform coefficients under a certain condition. Finally, for the remaining pseudo all zero blocks, machine learning is introduced and the decision is made through the fully connected neural network. Experimental results show that the proposed fast algorithm achieves up to 7.505% and 7.049% coding time saving under Low Delay B and Random Access configurations with only 0.470% and 0.578% performance loss on average, respectively. [ABSTRACT FROM AUTHOR]

It is an important task of modern traffic management for traffic police to inspect whether drivers wear seatbelts, and mobile law enforcement brings convenience to traffic police. However, due to the efficiency and memory constraints of mobile devices, the intelligent algorithm with significant parameters cannot be directly used on mobile devices. So, it is a challenge to detect driver and seatbelt by object detection algorithm on mobile. In order to solve this problem, we propose an efficient and lightweight model for seatbelt detection. First, we visualize the layers of SSD MobileNet V2 and delete the feature channels with low contribution and high similarity. Then we can get the pruned SlimSSDMV2 model for driver and seatbelt detection. Second, the LSD liner segment detection multipoint fitting algorithm is used for screening the undetected seatbelt area, which can further improve the detection performance of the model. We compare our model with other existing methods, and the experimental results demonstrate that our model performs better in practice. [ABSTRACT FROM AUTHOR]

In order to take advantage of the powerful advantages of heterogeneous devices and improve the robustness of the vehicle-mounted surround vision algorithm(VSVA), several key technologies of VSVA are improved in the paper. Firstly, computationally intensive tasks are calculated by heterogeneous Graphics Processing Unit(GPU), at the same time, so as to adapt to the VSVA, the memory model and computing model of GPU are optimized. Then a perspective transformation algorithm based on geometric constraints is proposed to improve the quality of the transformed image. Finally, an image alignment and fusion algorithm based on a calibration board is proposed, which reduces the complexity of the algorithm while ensuring the robustness of the image fusion algorithm. The paper compares the proposed algorithm with the traditional algorithm, the test results show that the proposed algorithm has good robustness and the overall performance of the VSVA is improved to 95.39%, the proposed algorithm can be widely used. [ABSTRACT FROM AUTHOR]

In the register detection of printing field, a new approach based on Zernike-CNNs is proposed. The edge feature of image is extracted by Zernike moments (ZMs), and a recursive algorithm of ZMs called Kintner method is derived. An improved convolutional neural networks (CNNs) are investigated to improve the accuracy of classification. Based on the classic convolutional neural network (CNN), the improved CNNs adopt parallel CNN to enhance local features, and adopt auxiliary classification part to modify classification layer weights. A printed image is trained with 7 × 400 samples and tested with 7 × 100 samples, and then the method in this paper is compared with other methods. In image processing, Zernike is compared with Sobel method, Laplacian of Gaussian (LoG) method, Smallest Univalue Segment Assimilating Nucleus (SUSAN) method, Finite Impusle Response (FIR) method, Multi-scale Morphological Gradient (MMG) method. In image classification, improved CNNs are compared with classical CNN. The experimental results show that Zernike-CNNs have the best performance, the mean square error (MSE) of the training samples reaches 0.0143, and the detection accuracy of training samples and test samples reached 91.43% and 94.85% respectively. The experiments reveal that Zernike-CNNs are a feasible approach for register detection. [ABSTRACT FROM AUTHOR]

With the continuous mining and gradual reduction of shallow deposits, deep prospecting has become a new global prospecting trend. In addition, with the development of artificial intelligence, deep learning provides a favorable means for geological big data analysis. This paper, researches the No. II Orebody of the Xiongcun deposit. First, based on previous research results and metallogenic regularity, prospecting information, namely, lithology, Au-Ag-Cu chemical elements and wall rock alteration is extracted, and the block model is established by combining the Kriging interpolation structure. Second, the datasets are divided into dataset I and dataset II according to "randomness" and "depth". Third, deep prospecting prediction models based on deep neural networks (DNN) and the convolutional neural networks (CNN) is constructed, and the model parameters are optimized. Finally, the models are applied to the deep prediction of the Xiongcun No. II Orebody. The results show that the accuracy rate and recall rate of the prediction model based on the DNN algorithm are 96.15% and 89.23%, respectively, and the AUC is 96.39%, which are higher values than those of the CNN algorithm, indicating that the performance of the prediction model based on the DNN algorithm is better. The accuracy of prediction model based on dataset I is higher than that of dataset II. The accuracy of deep metallogenic prediction based on the DNN algorithm is approximately 89%, that based on the CNN is approximately 87%, and that based on prospecting information method is approximately 61.27%. The prediction results of the DNN algorithm are relatively consistent in the spatial location and scale of the orebody. Therefore, based on the work done in this paper, it is feasible to use a deep learning method to carry out deep mineral prediction. [ABSTRACT FROM AUTHOR]

With the continuous mining and gradual reduction of shallow deposits, deep prospecting has become a new global prospecting trend. In addition, with the development of artificial intelligence, deep learning provides a favorable means for geological big data analysis. This paper, researches the No. II Orebody of the Xiongcun deposit. First, based on previous research results and metallogenic regularity, prospecting information, namely, lithology, Au-Ag-Cu chemical elements and wall rock alteration is extracted, and the block model is established by combining the Kriging interpolation structure. Second, the datasets are divided into dataset I and dataset II according to "randomness" and "depth". Third, deep prospecting prediction models based on deep neural networks (DNN) and the convolutional neural networks (CNN) is constructed, and the model parameters are optimized. Finally, the models are applied to the deep prediction of the Xiongcun No. II Orebody. The results show that the accuracy rate and recall rate of the prediction model based on the DNN algorithm are 96.15% and 89.23%, respectively, and the AUC is 96.39%, which are higher values than those of the CNN algorithm, indicating that the performance of the prediction model based on the DNN algorithm is better. The accuracy of prediction model based on dataset I is higher than that of dataset II. The accuracy of deep metallogenic prediction based on the DNN algorithm is approximately 89%, that based on the CNN is approximately 87%, and that based on prospecting information method is approximately 61.27%. The prediction results of the DNN algorithm are relatively consistent in the spatial location and scale of the orebody. Therefore, based on the work done in this paper, it is feasible to use a deep learning method to carry out deep mineral prediction. [ABSTRACT FROM AUTHOR]

Aiming at the problem of high cost and heavy preparation workload of indoor positioning technology, this paper proposes an ambient light-assisted smartphone-based pedestrian dead reckoning (PDR) indoor positioning technology. We propose a group weighted average algorithm to calculate the heading angle of pedestrians, which reduces the positioning error. During the PDR process, the smartphone continuously collects light intensity values, which are used to detect the light source location by using a light detection algorithm. Therefore, PDR has a self-correcting ability. The experimental results show that our method can effectively eliminate the cumulate error of traditional PDR and increase the operation distance. This paper has a certain contribution to the positioning of tunnels, underground shopping malls and other indoor places without sunlight interference. [ABSTRACT FROM AUTHOR]

The existing speech authentication algorithms use plaintext speech for hashing structure, and save the generated hashing sequence in the cloud, which is easy to cause hashing sequence leakage. At the same time, the robustness and matching accuracy of speech signal under complex noise are low, which will cause serious deviations in authentication. In order to solve the above problems, this paper proposes an encrypted speech perceptual hashing authentication algorithm based on improved two-dimensional Henon chaotic map (2D-Henon) encryption and harmonic product spectrum (HPS). Firstly, the algorithm sets the key for the improved 2D-Henon and transforms the original speech to encrypted speech. And then, feature extraction is performed on the encrypted speech signal to obtain the HPS feature matrix, and perform mapping and dimensionality reduction on the HPS matrix. Finally, the threshold is selected for binarization, and the perceptual hashing sequence is constructed. Experimental results show that the algorithm adopts encrypted speech perceptual hashing sequence, which improves the security of hashing sequence storage. The encrypted speech signal can reduce the amplitude of signal change, which has strong discrimination and robustness, and also has better matching accuracy against complex noise. [ABSTRACT FROM AUTHOR]

To solve the problems of low efficiency and poor resistance to attack, this paper proposes the multi-image encryption algorithm based on Haar wavelet transform and 3D shuffling scrambling. Taking advantage of the high computational efficiency of low-dimensional chaotic maps, this paper designs the dynamic pseudo-random sequence generator based on a dynamic chaotic library with three one-dimension chaotic maps and the roulette algorithm, which is highly associated with the plaintext image. To better scramble the image cube, this paper proposes a 3D shuffling scrambling algorithm, which divides the cube into one-dimensional vectors and scrambles the order of the one-dimensional vectors, then reorganizes the cube. To encrypt multiple images, first, reconstruct the images into an image cube and perform wavelet transformation on each layer of the cube. Then, use the 3D shuffling algorithm to scramble the low-frequency coefficient and reconstruct the cube with the scrambled low-frequency coefficient and high-frequency parts. Last, the chaotic matrix is XOR with each layer of the image cube. The algorithm can encrypt grayscale or color images of any size, which is flexible. In the simulation experiments, the algorithm has ideal ciphertext statistical characteristics, high running speed, and the ability of anti-attack, which is better than encryption algorithm in other references. [ABSTRACT FROM AUTHOR]

This paper proposes a collaborative filtering algorithm based on user group interest. A novel co-clustering method (BalClust) and various weighted non-negative matrix factorization algorithms are used in the proposed method. The BalClust method is used to divide the raw rating matrix into clusters, which are smaller than the original matrix. Then, the balance factor is introduced to consider the user weight and the item-based CF (collaborative filtering). To predict the rating of the unknown items in the cluster, the non-negative matrix factorization algorithm was used. The proposed method achieves higher predicting accuracy and efficiency on low dimensional and homogeneous sub-matrices, and the method also reduces the computational complexity by combining the user and item-based CF. Based on the proposed method, this paper proposed an incremental learning method to ensure data accuracy and timeliness to overcome the problem brought by data updates. The experimental results show the proposed methods outperformed traditional CF algorithms, and the completion time is reduced. [ABSTRACT FROM AUTHOR]

With the development of network technology, multimedia security has attracted extensive attention. Although chaotic system is widely used in this field because of its own characteristics, designing a more secure and efficient chaotic system and encryption algorithm has great application potential. In this paper, a new spatiotemporal chaos model, dynamic coupled perturbation map lattice (DCPML), is proposed. The variable function is introduced to replace the coupling input variable of spatiotemporal chaotic system, and the output of two-dimensional Logistic map is used as dynamic coupling coefficient and disturbance term to improve the pseudo randomness and chaos of the system. Through a series of analysis, it is proved that the model has excellent chaotic characteristics. The system is applied to chaotic encryption, and a new bit-level image encryption algorithm is designed. The algorithm breaks the correlation between pixels through bit-level confusion and diffusion operation. A large number of experiments and security analysis show that the algorithm can resist common attacks and has high security and robustness. [ABSTRACT FROM AUTHOR]

Information security is very important in the era of rapid development of science and technology. People often use multimedia to communicate in their daily life. Image plays an important role in multimedia communication, so it is urgent to protect image information. In order to improve the security of images in transmission, an image encryption algorithm using multi-base diffusion and a new four-dimensional chaotic system is designed in this paper. The algorithm decomposes each pixel value into multi-based. After decomposition, the coefficient matrix and base matrix are scrambled by FYTS using the sequence generated by the new four-dimensional chaotic system, and finally recombined to perform pixel-level and bit-level scrambling respectively. After simulation with MATLAB, it is obvious that the final encrypted image does not have the contour of the original image, which meets the requirements of encryption. Through histogram, entropy analysis, anti-differential attack analysis and other experimental results, it is proved that the proposed algorithm has high security. [ABSTRACT FROM AUTHOR]

This paper proposes a novel image encryption method based on coupled map lattice (CML) model. The main idea of this method is to solve the initial conditions of the CML system by combining the hash value of plaintext image with external keys. Then, the CML system is iterated to produce F intermediate chaotic sequences of f(x) map and X the chaotic sequences of CML system, respectively. The plaintext image is scrambled and diffused by using F chaotic sequences and the X chaotic sequences. Using the intermediate chaotic sequences F generated by CML system can effectively reduce the length of CML system and save the iteration time of chaotic sequence. Moreover, in the diffusion stage, forward diffusion and reverse diffusion are used to diffuse the image, which further enhances the security of the encryption algorithm. Experimental results indicate that the correlation coefficient of adjacent pixels in Lena's cipher image are 0.0005, 0.0065 and 0.0009 in the horizontal, vertical and diagonal directions respectively, and the information entropy is 7.9995, which is closer to ideal value 8. Experimental results and security analysis show that the proposed algorithm has high security performance and can resist various attacks. [ABSTRACT FROM AUTHOR]

Security and confidentiality are one of the main concerns when transmitting multimedia data over the Internet. To deal with the arising security issues, many advanced cryptographic algorithms have been proposed in the literature. The efficiency of these encryption algorithms varies according to the application area in which they are deployed. Although there are many objective evaluation metrics available in the literature to measure the performance of the cryptographic algorithms, however, to the best of our knowledge, no survey or review addresses the quality of the metrics for cryptography algorithms. In this paper, we have presented and analyzed the cryptography algorithms and performance evaluation metrics such as resource utilization measure, quality measure, and security measures to evaluate the security performance of different encryption algorithms for multimedia data. Extensive experiments have been carried out to evaluate the performance of the cryptography algorithms for multimedia data such as images and videos. We have also discussed and analyzed in detail the performance of evaluation measures used to validate the cryptography algorithms. Our findings highlight the strength of different cryptography algorithms and performance evaluation metrics. The analysis is expected to contribute to the design of a new and efficient model for multimedia applications. A detailed analysis of major attacks was also presented for each encryption scheme in tabular form. [ABSTRACT FROM AUTHOR]

To improve the resistance to geometric rotation attacks and to solve the problem of high computational complexity of existing watermarking algorithms, a novel efficient and robust zero watermarking algorithm in spatial domain is proposed. First, the central pixels of different channels of colorful host image are taken as the center of the circle, and the features of the image are constructed by the pixels covered by rings with different radius and width. The binary feature image is obtained by binary operation on the constructed image features. Then, the zero watermark image is constructed by XOR operation for the scrambled binary feature image and copyright watermark image encrypted Logistic chaotic algorithm. Finally, the zero watermark image is stored in the intellectual property database for copyright certification in possible copyright disputes. Experimental results show that the proposed algorithm is robust to geometric rotation attacks and common image processing attacks. Compared with similar zero watermarking schemes, the proposed algorithm has lower computational complexity and better robustness. [ABSTRACT FROM AUTHOR]

Authenticity in multimedia information retrieval system is one such parameter which is of utmost in demand. One of such parameters is the biometric based authentication system, as seen in case of human the system proves its importance in almost each and every field. But in case of animals it becomes quite complex to recognize them using unique biometric parameter. This paper presents a modified 10 layered DenseNet deep learning framework which is implemented for the identification of Humpback whale, in an online challenge conducted by the kaggle, using biometric parameter. The proposed dense model minimizes the traditional limitation of vanishing gradient problem and outperforms as compared with other methods exists in the literature. This paper also proposed an hand free flow algorithm for ROI segmentation. In addition, the proposed methodology also uses CLAHE as a preprocessing method and focused on the feature re-usability with an optimized number of parameters used for recognition. During the experimentation, validation of the proposed model is found satisfactory with an accuracy of 93.21%. The test set results are provided by the kaggle with an AUC-ROC result over public and private scoreboard with an accuracy of 94.47% and 92.54% respectively. The comparative result with other deep methodologies suggests that the highest accuracy gained is 95.55% by Se-ResNet50 by some other learderboard and proposed model is found at second place. [ABSTRACT FROM AUTHOR]

In order to improve the contrast, clarity and color fidelity of low illumination image, reduce the negative impact of image degradation, in this paper, we propose an image clarity scheme based on multi-image multi-parameter and multi-scale fusion. The scheme is based on the original image, according to our algorithm to generate the optimal exposure image, overexposure image and variational Retinex image. Secondly, the four images are taken as input seperately, and each input is decomposed into contrast component, texture structure component and brightness component. Finally, the extracted contrast component, texture structure component and brightness component are fused respectively according to the weight to build the image enhancement model, then get the enhanced image. Comparing different low illumination images, the experimental results show that the color cast of the proposed method is reduced by 47.8%, the contrast value are increased by 53%, the standard deviation is increased by more than 1.18 times, and the average gradient is increased by more than 4 times compared with the original image. The comprehensive performance is better than other comparison algorithms. The method in this paper has been well improved in image enhancement and detail preservation. It effectively improves the color fidelity, contrast and detail clarity of the image under low illumination. The enhanced visual effect and visibility of the image are obviously improved and more real and natural. [ABSTRACT FROM AUTHOR]

Versatile Video Coding (VVC/H.266) expands the intra-frame prediction modes from 35 of H.265 to 67, which not only improves the prediction accuracy and coding efficiency of the encoder but also increases its encoding complexity. In response to this problem, this paper proposes a support vector machine-based fast planar prediction mode decision algorithm for VVC/H.266, which is used to quickly determine the Planar or non-Planar prediction modes during intra coding, avoiding the rate-distortion optimization calculation of multiple intra prediction modes, and thus reducing the encoding time. The contributions of this paper are: 1) It proposes a novel feature based on Statistical Oriented Gradient (SOG) to extract the feature information of the coding block; 2) Based on the SOG feature, a support vector machine-based method is proposed to make the decision between the Planar and non-Planar modes for intra coding of VVC/H.266. The proposed algorithm has been integrated into the VVC/H.266 standard coder, VTM5.0. The experimental results show that, compared with VTM5.0, the encoding time of the proposed algorithm is reduced by 18.0% on average, while the Bjøntegaard-Delta -Rate (BD-RATE) increased by only 1.3%. [ABSTRACT FROM AUTHOR]

BeiDou-3 satellites have short-message communication capabilities, which are not available in other navigation systems. However, the information transmission capacities of BeiDou-3 satellites are limited, which means that the string length of a single transmission is restricted. To improve transmission performance, this paper proposes a lossy data compression algorithm to transmit ship data. First, due to the large amount of measurement noise in the ship data and the difficulty of adaptively adjusting the compression threshold of the swinging door trend, a dynamic threshold adjustment strategy based on model-free adaptive control is designed. Furthermore, because the performance of the strategy depends on the initial value, chaos multilayer particle swarm optimization is employed to optimize the strategy and achieve better generalization and convergence performance. In addition, this paper introduces an experimental system and uses real ship motor vibration and electromagnetic radiation data to verify the compression algorithm. The experimental results indicate that the proposed algorithm displays reasonable performance. Presently, the system designed in this paper has been operating stably for more than six months. [ABSTRACT FROM AUTHOR]

In this paper, a new four-dimensional hyperchaotic system that has better unpredictability, more complex dynamic behavior, and a larger key space is proposed. The initial chaos value is calculated by using the Secure Hash Algorithm-512 (SHA-512) function, and the hyperchaotic sequence is then used as the key stream to encrypt the original image through pixel scrambling and pixel diffusion. The experimental results show that the image encryption algorithm has a strong ability to resist exhaustive attacks, statistical attacks and plaintext attacks. [ABSTRACT FROM AUTHOR]

This paper proposes a new image encryption algorithm using novel block-level scrambling scheme and bit-level filer diffusion technology. In order to increase the diversity of pseudo-random number generator (PRNG) and improve the security of encryption system, a new Josephus travel scheme is proposed in this paper. Then, based on the novel Josephus travel method, a novel block-level scrambling scheme and bit-vector diffusion operation are proposed to improve the encryption efficiency of the algorithm. Numerical simulation and comparison experiments shows that the proposed encryption algorithm has high security and faster encryption speed, and can resist various attacks. [ABSTRACT FROM AUTHOR]