This work introduces an unsupervised framework for video anomaly detection, leveraging a hybrid deep learning model that combines a vision transformer (ViT) with a convolutional spatiotemporal relationship (STR) attention block. The proposed model addresses the challenges of anomaly detection in video surveillance by capturing both local and global relationships within video frames, a task that traditional convolutional neural networks (CNNs) often struggle with due to their localized field of view. We have utilized a pre-trained ViT as an encoder for feature extraction, which is then processed by the STR attention block to enhance the detection of spatiotemporal relationships among objects in videos. The novelty of this work is utilizing the ViT with the STR attention to detect video anomalies effectively in large and heterogeneous datasets, an important thing given the diverse environments and scenarios encountered in real-world surveillance. The framework was evaluated on three benchmark datasets, i.e., the UCSD-Ped2, CHUCK Avenue, and ShanghaiTech. This demonstrates the model’s superior performance in detecting anomalies compared to state-of-the-art methods, showcasing its potential to significantly enhance automated video surveillance systems by achieving area under the receiver operating characteristic curve (AUC ROC) values of 95.6, 86.8, and 82.1. To show the effectiveness of the proposed framework in detecting anomalies in extra-large datasets, we trained the model on a subset of the huge contemporary CHAD dataset that contains over 1 million frames, achieving AUC ROC values of 71.8 and 64.2 for CHAD-Cam 1 and CHAD-Cam 2, respectively, which outperforms the state-of-the-art techniques.
Personal identification systems based on electroencephalographic (EEG) signals have their own strengths and limitations. The stability of EEG signals strongly affects such systems. The human emotional state is one of the important factors that affects EEG signals’ stability. Stress is a major emotional state that affects individuals’ capability to perform day-to-day tasks. The main objective of this work is to study the effect of mental and emotional stress on such systems. Two experiments have been performed. In the first, we used hand-crafted features (time domain, frequency domain, and non-linear features), followed by a machine learning classifier. In the second, raw EEG signals were used as an input for the deep learning approaches. Different types of mental and emotional stress have been examined using two datasets, SAM 40 and DEAP. The proposed experiments proved that performing enrollment in a relaxed or calm state and identification in a stressed state have a negative effect on the identification system’s performance. The best achieved accuracy for the DEAP dataset was in the calm state and in the stressed state. For the SAM 40 dataset, the best accuracy was , , , and for the relaxed state and stress caused by identifying mirror images, the Stroop color-word test, and solving arithmetic operations, respectively.
This paper proposes an efficient lightweight deep spatial residual autoencoder (SRAE) model to detect anomalous events in video surveillance systems. A lightweight network is essential in real-time situations where time is critical. Moreover, it could be deployed on low-resource devices like embedded systems or mobile devices. This makes it a very useful option for real-world situations where there may be a shortage of resources. The proposed network is composed of a 3-layer residual encoder-decoder architecture that is adopted to acquire the salient spatial characteristics representative of normal events in videos. Then, the reconstruction loss is used to find abnormalities, where normal frames are recreated well with a low reconstruction loss, and abnormal ones are found as the opposite. The model's efficiency is tested by two benchmark datasets, the UCSD Pedestrian2 and the CUHK Avenue, achieving AUC ≈ 95% and 81% for the two datasets, respectively. Hence, its performance proves to be comparable with the state-of-the-art models.
In this paper, we propose an eHealth system based on Health-level7 – Clinical Document Architecture targeting the public healthcare sector in Egypt. The System uses HL7, which is an international and well-recognized standard, to build electronic medical record while providing complete privacy and utmost security of patients’ data. A novel property, that the system provides, is the online video consultation service which has become mandatory these days due to the widespread of COVID-19. Additionally, the system ensures the credibility of the physician’s identity and practice through networking with the ministry of health and physicians’ syndicate. The proposed system could also be accessible on an Android-based cell phone with all features fully functioning. Testing of the system’s functionality and acceptability was conducted. It was noted that the system was well accepted by both parties; patient and physician. A comparison was carried out with other e-health systems highlighting the points of strength that the proposed system presents. The system has an extendable infrastructure that makes it flexible to integrate later with other systems.
Real-time systems must avoid hazardous situations. To achieve this, their functionality should be investigated under time constraints. In this paper, a modeling based on Hierarchical Communicating Real-time State Machine (H- CRSM) and analysis methodology is proposed to examine statically ANSI-C code to obtain the hazardous events in the input system. A hazardous event equation is taken as input to the proposed system. The output is a list of hazardous scenarios. A path in the code showing the cause of the undesirable event is associated with each hazardous scenario. The strength of the proposed methodology is that the process of hazardous situations detection does not require running the ANSI-C program multiple times with different input values. It also focuses on analyzing the software level of the life cycle. Most of the verification tools check the system level. The system level may be bug-free but the software level may not.