Optimizing YOLOv8n using hOSVD for efficient object detection: towards real-time video surveillance applications

This work presents a video-based sequence synchronization algorithm to be used in real-time video surveillance applications. The signals are aligned based on an online dynamic time warping approach that uses only video content information. The algorithm was tested in the alignment of reference and target videos acquired in a cluttered industrial environment with a moving camera. During each recording, the camera follows approximately the same trajectory but its speed along the trajectory may differ among recordings, thus generating time warping between the videos. The results shows that, when considering both alignment error and computational complexity, the use of the mean squared error as similarity metric leads to the best performance among metrics such as the SSIM, absolute error or even sophisticated matching techniques such as histograms of oriented gradients. The proposed method was also tested in a real-time application. Results show that the instantaneous alignment computation of a target frame can be performed successfully, at the cost of a three-fold increase in the error when compared to the case when latency is allowed.

Uses of video surveillance (VS) have exponentially increased using the internet as a platform. Security concerns related to the transmission of videos must be addressed. Even though the majority of current encryption standards are effective, it is computationally demanding and unsuitable for real-time VS applications. Lightweight crypto-encryption standards are necessary to meet real-time needs. This paper provides a thorough analysis of various encryption standards to protect video data and evaluates vulnerability under various attacks. A Lightweight Fast Security Standard (LFSS) is proposed to meet real-time requirements by adopting dual key fusion (KF) approach. Entropy and crypto weights are quantitatively compared with proposed and chosen lightweight ciphers for surveillance videos based on NCPR, UACI and attacks into account. The proposed LFSS method offers a good NCPR of 99.5806 and offers only 40% and 3% average execution time over Secure Internet of things (SIT) and Secure Force (SF) security standards, respectively.

Human Activity analysis is one of the most captivating and important open problem in automated video surveillance community. In recent years, most of the analysis of human activities/behavior is carried out using computer vision & pattern recognition techniques and has paved the way for amalgamation of various such fields. This paper gives an assessment of this new development, proposed to analyze the impact of theoretical and practical contributions of two well known techniques such as HOG & STIP made along with the performances of different distance measures on two benchmarking databases - KTH & Weizman Database. Some crucial experiments are conducted along with the generalized evaluation techniques to give fair and integrated analysis. We also highlight significant challenges involved in utilizing these techniques in real time video surveillance applications. At the end of this paper, inferences are made in regard to future research directions in human activity analysis.

Real-time video surveillance, especially CCTV systems, requires fast and accurate face detection. Object detection models with slow inference times are ineffective in real-time. This study addresses this challenge by improving the inference speed of the YOLOv8 model, a leading object detection framework known for its accuracy and speed. We focus on pruning the model's architecture, particularly the P5 head section, which detects larger objects. According to Bochkovskiy's 2020 research, this modification enhances the model's performance specifically for medium and small objects in CCTV footage. The standard YOLOv8 model and its modified version were compared for inference time, mean Average Precision (mAP), and model weight. The pruned YOLOv8 model cuts inference time by 15.56%, from 4.5 ms to 3.8 ms, and reduces model weight. The advantages mentioned above are offset by a 1.6% decrease in mean average precision. This research advances object detection technology by demonstrating architectural modifications' efficacy. These changes make the model faster and lighter, making it suitable for real-time surveillance. The accuracy trade-off is slight. The implications of these findings are crucial for implementing efficient object detection systems in CCTV surveillance. These findings also lay the groundwork for future research to improve such systems' speed-accuracy trade-off.

The construction of a model of the background of a scene still remains as a challenging task in video surveillance systems, in particular for moving cameras. This work presents a novel approach for constructing a panoramic background model based on competitive learning neural networks and a subsequent piecewise linear interpolation by Delaunay triangulation. The approach can handle arbitrary camera directions and zooms for a Pan-Tilt-Zoom (PTZ) camera-based surveillance system. After testing the proposed approach on several indoor sequences, the results demonstrate that the proposed method is effective and suitable to use for real-time video surveillance applications.

Now a day, video surveillance has gained the most attention in real-time applications. Because of the improvement in domain of machine learning, various methodologies have been designed for multi-object detection. The real-time surveillance systems are the need of various organization as security; hence, this area of research has important. This paper introduces a new methodology for Moving object detection. In the area of real-time video surveillance systems, multi-object detection and object tracking are an important step in video processing. The moving system is critical for many of the application such as operational robot, surveillance systems in military. These real-time surveillance systems are more useful for intelligence gathering, security issues, and many personal needs. Different techniques are used for this task and research is vastly done to make this system automated and to make it reliable. In this research, subjective quality assessment of object detection and object tracking is discussed in detail. In the proposed system, the background subtraction is done from the clean original image and use of CNN with this. We have applied background subtraction method with different filtering algorithms. We have implemented and tried MoD2 and KNN algorithm on the test video and fetched the foreground image. After this, we have applied CNN to track the object. After tracking the objects, we have found out the accuracy with background subtraction method and without background subtraction method. The proposed system improved by approximately 2% as compared to existing techniques.

- Managing crowd safety during the Kumbh Mela presents unprecedented logistical challenges due to its scale,density, and dynamic environment. This work introduces Kumbh-Connect, a real-time intelligent surveillance system that leverages computer vision and machine learning to enhance public safety.The framework combines YOLOv11n for person and weapon detection, ByteTrack for movement analysis, and DeepFace form facial recognition of missing individuals. A FastAPI backend processes video streams from CCTV feeds, while WebSockets enable instant alerts to security personnel via a mobile application.Testing on simulated event environments shows 90% accuracy in crowd estimation and 92% in facial matching. Designed for scalability, KumbhConnect offers a proactive approach to crowd control, reducing response time and improving situational awareness during large-scale gatherings. Index Terms—Kumbh Mela, Crowd Monitoring, Object Detection, Facial Recognition, Real-Time Surveillance, YOLO.

Low complexity object detection and tracking with inter-layer graph mapping and intra-layer graph refinement in H.264/SVC bitstreams

Adopting Real-Time Surveillance Dashboards as a Component of an Enterprisewide Medication Safety Strategy

This research focuses on intelligent unmanned aerial vehicle (UAV)-based, real-time video surveillance to ensure better monitoring and security of remote locations over 4G-LTE cellular networks by maximizing end-user quality of experience (QoE). We propose a novel server-based crosslayer rate-adaptive scheme (SCRAS) for real-time video surveillance over 4G-LTE networks using UAVs. Our key contributions are: (1) In SCRAS, mobile UAVs having preprogrammed flight co-ordinates act as servers, streaming real-time video towards a remote client; (2) server-side video rate adaptation occurs in 4G-LTE based on the physical characteristics of the received signal conditions due to variations in the wireless channel and handovers; (3) SCRAS is fully automated and independent of client assistance for rate adaptation, as it is intended for real-time, mission-critical surveillance applications; (4) SCRAS ensures that during rate adaptation, the current video frame should not be damaged by completing the current group of packet (GoP) before adaptation. Our simulations in NS3 provide credible evidence that SCRAS outperforms recently proposed schemes in providing better QoE for real-time, rate-adaptive, video surveillance over 4G-LTE under varying channel quality and frequent handovers that occur during flight by UAVs.

Objective Some countries have implemented systems to monitor suicides in real-time. These systems differ because of the various ways in which suicides are identified and recorded. The main objective of this study was to conduct an international comparison of major real-time suicide mortality surveillance systems to identify joint strengths, challenges, and differences, and thereby inform best-practice criteria at local, national, and international levels. Methods Five major real-time suicide mortality surveillance systems of various coverage levels were identified and selected for review via an internet-based scoping exercise and prior knowledge of existing systems. Key information including the system components and practices was collated from those organizations that developed and operate each system using a structured template. The information was narratively and critically synthesized to determine similarities and differences between the systems. Results The comparative review of the five established real-time suicide surveillance systems revealed more commonalities than differences overall. Commonalities included rapid, routine surveillance based on minimal, provisional data to facilitate timely intervention and postvention efforts. Identified differences include the timeliness of case submission and system infrastructure. Conclusion The recommended criteria could promote replicable components and practices in real-time suicide surveillance while offering flexibility in adapting to regional/local circumstances and resource availability. HIGHLIGHTS Evidence-informed recommendations for current best practice in real-time suicide surveillance. Proposed comprehensive framework can be adapted based on available resources and capacity. Real-time suicide mortality data facilitates rapid data-driven decision-making in suicide prevention.

This paper presents a real-time surveillance system for detecting and tracking people, which takes full advantage of local texture patterns, under a stationary monocular camera. A novel center-symmetric scale invariant local ternary pattern feature is put forward to combine with pattern kernel density estimation for building a pixel-level-based background model. The background model is then used to detect moving foreground objects on every newly captured frame. A variant of a fast human detector that utilizes local texture patterns is adopted to look for human objects from the foreground regions, and it is assisted by a head detector, which is proposed to find in advance the candidate locations of human, to reduce computational costs. Each human object is given a unique identity and is represented by a spatio-color-texture object model. The real-time performance of tracking is achieved by a fast mean-shift algorithm coupled with several efficient occlusion-handling techniques. Experiments on challenging video sequences show that the proposed surveillance system can run in real-time and is quite robust in segmenting and tracking people in complex environments that include appearance changes, abrupt motion, occlusions, illumination variations and clutter.

Bus stops in Riobamba had significant deficiencies in safety, accessibility, and comfort, which limited the effective use of public transport and affected the urban mobility of the population. Improving these conditions was crucial to promote sustainable, inclusive and safe mobility in the city. This study was quantitative and descriptive, based on 420 user surveys and the direct observation of 140 stops, complemented with georeferencing and comparative review of specialized literature. The findings showed that most of the stops lacked adequate lighting, shelter, signage and universal access, with 68% of users perceiving low safety. The most in-demand technologies included real-time information systems (72%) and video surveillance (65%). The proposed model of smart stops will improve accessibility, safety and comfort for users, encouraging greater use of public transport. By addressing the main infrastructure and technology gaps, the intervention contributed to inclusive and safe urban mobility, directly contributing to Sustainable Development Goal 11 and offering a replicable framework for other medium-sized cities seeking to optimize their public transport systems.

A number of real-time video traffic monitor systems are currently being developed by the Texas Transportation Institute (TTI), Texas A&M University System, to demonstrate how new Intelligent Transportation Systems (ITS) technologies, such as wireless communication, ISDN communication, standard phone-line video compression technology, real-time video surveillance, and video image processing, can be better integrated. These systems were used to illustrate how video- based traffic monitoring and resource sharing can be better achieved to improve system efficiency and increase operational safety by operating agencies anywhere in the world.

This paper presents a system for the estimation of vehicle's trajectories in the context of real-time traffic video surveillance applications. The scene characteristics (regions of interest, scene depth and the circulation lanes) are assumed to be available. It is used together with a vehicle geometrical model during the vehicle detection, tracking and trajectory estimation phases to improve the robustness against the perspective and occlusion effects. It proposes a similarity criterion and a prediction phase based on the estimated typical trajectories and the circulation lanes. Experiments show that most of the time, the trajectories can be correctly estimated in real time.