Abnormal crowd behaviors in high density situations can pose great danger to public safety. Despite the extensive installation of closed-circuit television (CCTV) cameras, it is still difficult to achieve real-time alerts and automated responses from current systems. Two major breakthroughs have been reported in this research. Firstly, a spatial-temporal texture extraction algorithm is developed. This algorithm is able to effectively extract video textures with abundant crowd motion details. It is through adopting Gabor-filtered textures with the highest information entropy values. Secondly, a novel scheme for defining crowd motion patterns (signatures) is devised to identify abnormal behaviors in the crowd by employing an enhanced gray level co-occurrence matrix model. In the experiments, various classic classifiers are utilized to benchmark the performance of the proposed method. The results obtained exhibit detection and accuracy rates which are, overall, superior to other techniques.

The research proposes an innovated structure of the density map-based crowd counting network augmented by information entropy. The network comprises of a front-end network to extract features and a back-end network to generate density maps. In order to validate the assumption that the entropy can boost the accuracy of density map generation, a multi-scale entropy map extraction process is imported into the front-end network along with a fine-tuned convolutional feature extraction process, In the back-end network, extracted features are decoded into the density map with a multi-column dilated convolution network. Finally, the decoded density map can be mapped as the estimated counting number. Experimental results indicate that the devised network is capable of accurately estimating the count in extremely high crowd density. Compared to similar structured networks which don’t adapt entropy feature, the proposed network exhibits higher performance. This result proves the feature of information entropy is capable of enhancing the efficiency of density map-based crowd counting approaches.

Traditional detect and count strategy can’t well handle the extremely crowded footage in computer vision-based counting task. In recent years, deep learning approaches have been widely explored to tackle this challenge. By regressing visual features to density map, the total crowd number can be predicted while avoids the detection of their actual positions. Efforts of improving performance distribute at various phases of the detecting pipeline, such as feature extraction and eliminating deviation of regressed density map etc. In this article, we conduct a thorough review on the most representative and state-of-the-art techniques. The efforts are systematically categorized into three topics: the evolving of front-end network, the handling of unbalanced density map prediction, and the selection of loss function. After the evaluation of most significant techniques, innovations of the state-of-the-art are inspected in detail to analyze specific reasons to achieve high performances. As conclusion, possible directions of enhancement are discussed to provide insights of future research.

Abnormal behaviour detection algorithm needs to conduct behaviour analysis on the basis of continuous video inclination tracking, and the robustness of the algorithm is reduced for the occlusion of moving targets, the occlusion of the environment, and the movement of targets with the same colour. For this reason, the optical flow information between RGB (red, green, and blue) images and video frames is used as the input of the network in view of group behaviour. Then, the direction, velocity, acceleration, and energy of the crowd were weighted and fused into a global optical flow descriptor. At the same time, the crowd trajectory map is extracted from the original image of a single frame. Following, in order to realize the detection of large displacement moving target and solve the problem that the traditional optical flow algorithm is only suitable for the detection of displacement moving target, a video abnormal behaviour detection algorithm based on the double‐flow convolutional neural network is proposed. The network uses two network branches to learn spatial dimension information and temporal dimension information, respectively, and uses short‐ and long‐time neural network to model the dependency relationship between long‐time video frames, so as to obtain the final behaviour classification results. Simulation test results show that the proposed method can achieve good recognition effect on multiple datasets, and the performance of abnormal behaviour detection can be significantly improved by using interframe motion information.

Techniques of abnormal behavior recognition enhance public safety by analyzing the vision-based biometric features of the individual. However, main-stream deep learning-based approaches heavily rely on hardware computility, which is difficult to deploy on edge devices. This defect harms their practical integrity as a biometric identification system and the potential for commercial application. We introduce a novel lightweight framework for edge-based real-time anomaly detection by combining lightweight MobileInst architecture with temporal modules namely MobLSTM and MobGRU. The MobLSTM improves capability of detecting gradual or nuanced abnormal activities by incorporating temporal LSTM model. And the MobGRU simplifies information flow from backbone to minimize the computational burden. The Query-Passing Mechanism is also adapted to further reduce computation redundancy by reusing queries obtained from consecutive frames. The approach is evaluated on UCF-Crime dataset, a benchmark widely used in abnormal behavior detection. Experiment results indicate MobLSTM yields a frame-level anomaly detection accuracy of around 92.82% at 90.8 FPS and MobGRU with the accuracy of 89.59% at 92.3 FPS. These findings further highlight the practicality of deploying high performance anomaly detection systems in resourceconstrained environments, presenting a scalable solution for real-world surveillance, and demonstrating significant implications in the improvement of public safety via conveniently available real-time biometric identification systems.

Deep learning models perform exceptionally well in image processing; however, the increasing number of parameters and computational burdens make it difficult to apply these models on mobile devices. To address this issue, a widely adopted approach is using knowledge distillation to compress and optimize deep learning models. Traditional knowledge distillation trains a student network with soft targets from a teacher network. This paper proposes a Novel Dual Knowledge Distillation (NDKD) network by integrating Convolutional Block Attention Module (CBAM) distillation and feature distillation. Specifically, a new CBAM module and a hash layer are first added to improve the teacher-student network. Next, attention distillation loss is computed to transfer CBAM knowledge from the teacher module to the student network to enhance the student network's ability to capture low-level visual features. Finally, feature distillation loss is constructed by hashing layers to transfer semantic knowledge from the teacher network to the student network to enhance the student network's high-level semantic ability. Our method is compared with other knowledge distillation methods on benchmarking datasets including the CIFAR-10, SVNM, EMNIST, and Tire Pattern Image dataset CIIP-TPID. Experimental results show that our method outperforms other methods on adapted datasets.