Live Video Streaming Research Articles

IoT-Based Metal Detector Robot with Wireless Surveillance

Abstract: This project presents the design and development of an IoT-based metal detector robot equipped with wireless surveillance capabilities. The robot is designed to autonomously navigate its environment while detecting metallic objects underground, with real-time monitoring provided through a wireless camera system. The core components of the system include a metal detector sensor, an ESP8266 NodeMCU for wireless communication, an L298 motor driver for controlling the robot’s movement, and a surveillance camera for live video streaming. The system is powered by 3.7V battery cells, ensuring mobility and operational autonomy.The metal detector sensor scans for metallic objects, and upon detection, an alert is generated. Simultaneously, a wireless camera transmits real-time footage, allowing remote monitoring and control via a web interface or mobile device. The robot's movement is controlled through a second ESP8266 module, enabling wireless command inputs for navigation.This robot is envisioned for use in security applications, such as detecting landmines or hazardous metallic objects in sensitive areas, as well as for archaeological and construction projects. The integration of IoT technology enhances the robot’s operational range and effectiveness, providing a scalable solution for remote metal detection and surveillance.The integration of IoT technology significantly enhances the system's operational efficiency and range, enabling seamless remote control and monitoring. This innovation not only improves safety in hazardous conditions but also increases the effectiveness of metal detection and surveillance tasks in diverse environments

VidBlock: A Web3.0-Enabled Decentralized Blockchain Architecture for Live Video Streaming

In the digital era, the demand for real-time streaming services highlights the scalability, data sovereignty, and privacy limitations of traditional centralized systems. VidBlock introduces a novel decentralized blockchain architecture that leverages the blockchain’s immutable and transparent characteristics along with direct communication capabilities. This ecosystem revolutionizes content delivery and storage, ensuring high data integrity and user trust. VidBlock’s architecture emphasizes serverless operation, aligning with the principles of decentralization to enhance efficiency and reduce costs. Our contributions include decentralized data management, user-controlled privacy, cost reduction through a serverless architecture, and improved global accessibility. Experiments show that VidBlock is superior in reducing latency and utilizing bandwidth, demonstrating its potential to redefine live video streaming in the Web3.0 era.

Going Live: A Look at Memory, Enjoyment, Perceived Learning and Credibility, Across Live Streaming and Non-Live Short-Form Video Viewing

Short-form videos, often found on TikTok but available across social media, are increasing in both popularity and in the frequency that they are sought out for informational purposes. Short-form videos can be delivered via live streaming or via pre-recorded messages, both of which provide different affordances to users. We explore two of these affordances: The synchronous communication between viewers, creators, and moderators permitted by live streaming short-form videos, and the presence of visuals in videos, which provide complementary information via a second perceptual processing stream. An experiment with an undergraduate student sample (N = 412) was conducted. Data demonstrates a clear pattern of live streaming (with its synchronous communication affordances) resulting in greater enjoyment, perceived credibility, perceived learning, and recall memory, whereas the presence of visuals does not influence most of these outcomes. Implications are discussed.

IMMEDIATE OBJECT SPOTTING

Object Spotting has become one of the most essential and challenging tasks in computer vision, particularly in the context of deep learning advancements. This paper reviews the evolution of object detection techniques, highlighting how deep neural networks have transformed the field. We classify the state-of-the-art detection algorithms into three main categories: anchor-based, anchor-free, and transformer-based detectors, each with distinct methodologies for identifying objects in images. This survey includes a comparison of key convolutional neural network (CNN) architectures used for object detection, evaluating their speed-accuracy trade-offs, quality metrics, and training strategies. We discuss the strengths and limitations of these models and offer insights into the future directions of research in object spotting. In parallel, immediate object spotting using OpenCV has gained prominence for its ability to identify and localize objects within live video streams, particularly by leveraging pre-trained models like YOLOv3. This approach processes each video frame in real time, with preprocessing steps such as resizing and normalization to ensure compatibility with the deep learning model. YOLOv3’s efficiency enables it to detect and classify objects quickly, drawing bounding boxes around detected objects and displaying them in real time. This project highlights how combining OpenCV with deep learning models can achieve effective, scalable object spotting while ensuring detection accuracy. The paper also explores the challenges and potential future advancements in improving detection accuracy, robustness, and compatibility with modern hardware systems for diverse applications. Keywords:Immediate object spotting, OpenCV, YOLOv3, Deep learning models, Live video stream, Pre- processing (resizing, normalization), Object identification and localization, Bounding boxes, Detection accuracy, System compatibility

Real-Time People Counting System

This paper introduces an advanced system for real-time people counting, utilizing live video streams or IP cameras. The core of this system is the Single Shot Multibox Detector (SSD) framework, enhanced with MobileNet architecture to ensure precise object detection. To continuously track detected individuals, we employ a centroid tracking algorithm. The system's design prioritizes scalability, making it ideal for various business applications, such as monitoring foot traffic in stores, buildings, and shopping malls. A key feature of our system is its ability to send immediate alerts when the number of people exceeds a set threshold. This functionality is crucial for managing crowd control and adhering to safety protocols, particularly in scenarios like the COVID-19 pandemic. We have optimized the system’s performance with threading, which improves its responsiveness and efficiency. By integrating advanced machine learning techniques with practical business applications, our real-time people counting system offers a dependable and scalable solution for contemporary crowd management challenges. Features like real-time alerts, automatic scheduling, and comprehensive data logging highlight its effectiveness in enhancing operational efficiency and safety across various settings.

Facial Recognition Using Video Streaming

Facial recognition has become an essential technology in numerous applications, including security, surveillance, and human-computer interaction. With the advent of real-time video streaming, facial recognition systems are now tasked with handling the continuous flow of visual data. This paper explores the integration of facial recognition algorithms into video streaming systems, investigating the challenges, methodologies, and applications of real-time face detection and identification. We present a framework for implementing facial recognition in live video streams, assess the performance of different algorithms, and highlight areas for future improvement, including handling large datasets, optimizing computation, and ensuring privacy.

Performance Analysis of Underwater Radiofrequency Communication in Seawater: An Experimental Study

Communication with the underwater vehicles during their tasks is one of the most important issues. The need for real-time data transfer raises the necessity of developing communication systems. Conventional underwater communication systems, such as acoustic systems, cannot satisfy applications that need a high transmission data rate. In this study, we investigate the radio frequency communication system in seawater, which is crucial for real-time data transfer with underwater vehicles. The experiments were in a water tank full of seawater and a real environment in the ocean. Three types of antennae were used: loop antenna, wire antenna, and helical antenna. An Autonomous Underwater Vehicle (AUV) is used as a transmitter to measure the transmission rate as a function of distance. The helical antenna showed better performance regarding the coverage area. Furthermore, the AUV could move freely within the helical and capture live video streaming successfully. This investigation underscores the potential of radio frequency communication systems for enhancing underwater vehicle operations, offering promising avenues for future research and practical implementation.

Incentivized BiLSTM with Triplet Attention for Predecting Congestion in Network Traffic

In the world of online live video streaming, network traffic congestion poses a serious risk to user experience and service quality. Several factors contributing to the congestion include insufficient bandwidth, an increase in user demand, as well as ineffective data routing. The incapacity of the prediction algorithms to Adjust to modifications in network circumstances and challenges in detecting long-range relationships may restrict their ability to effectively estimate congestion in dynamic contexts when it comes to online streaming video. In order to overcome these constraints, this study created an Incentivized-RF-Triplet ASTM (“Incentivized learning-based triplet attention enabled rat fierce hunting optimized Bidirectional Long Short-Term Memory) for network traffic congestion prediction in online streaming video. A reward” scheme built into the Incentivized learning mechanism pushes the algorithm to prioritize online live video streaming congestion prediction. The sequential structure of network traffic data is handled by the BiLSTM architecture, that is well-known for capturing temporal dependencies. By utilizing the triplet Attention method, the model is better able to identify relevant regions within the input data as well as identify congestion patterns more successfully. The RFSO method incorporates social behavior with selection and searching qualities to further improve the classifier's parameters. Therefore, the characteristics of the model can be tuned more effectively and robustly, improving its effectiveness in congestion detection. The outcomes of the experiment show how well the Incentivized-RF-Triplet ASTM technique works to precisely estimate traffic congestion for the Darpa99week1 dataset. 95.97% accuracy, 96.08% specificity, and 0.22 mean square error are reported.

Leveraging openCV and image processing for crime identification through facial recognition

To maintain the safety and security of communities, law enforcement agencies provide vital functions such as crime prevention and identification. Identifying people with a criminal record efficiently and accurately is one of the biggest hurdles in this field. This study delves into a fresh strategy for taking on this problem by making use of state-of-the-art image processing methods and machine learning algorithms. The project's main goal is to identify convicted felons by utilizing OpenCV, a robust open-source computer vision toolkit, and different facial recognition algorithms. By giving them a tool that guarantees a faster reaction rate and improved accuracy in suspect identification, the system is aimed to boost the capacities of law enforcement organizations. This project is based on the tenet of criminal psychology, which states that criminals are prone to committing new crimes. The system's goal is to detect and track repeat criminals both before and after they commit crimes by making use of this information. The method works by taking pictures of people who are known to be criminals and putting them in a database. From there, they can be compared to either live video streams or static photographs taken by surveillance cameras. In order to analyze and identify facial characteristics effectively, algorithms like LBPH and the Haar cascade classifier are also used. The system's robustness and efficiency are guaranteed by integrating these approaches into the OpenCV framework. This research makes a valuable contribution to the field of crime prevention by offering law enforcement authorities a dependable and scalable alternative. This system provides a streamlined interface for storing, retrieving, and matching facial photos. It was built using Python 3.5 and uses SQLite to manage the database of offenders' details. Proactively monitoring persons with a criminal record is made easier and faster with the suggested method. Suspect identification is also made easier. This research showcases the possibility of integrating machine learning and image processing to enhance public safety and crime prevention initiatives. It is an early step towards more complete video surveillance systems.

Exploring The Celebrity Effect on Live-Streaming E-Commerce: A Comprehensive SWOT Analysis of Influencer Marketing Strategies

In recent years, with the rapid development of the Internet and mobile technology, the landscape of business and marketing has changed dramatically, and particularly notable is the rise of live-streaming e-commerce. This new e-commerce model combines e-commerce with live video streaming to enhance the shopping experience through product display, promotion and sales by anchors. In this model, the celebrity effect, which takes advantage of celebrities' large fan bases and strong influence, plays an important role in driving consumer behaviour. This paper explores the impact of the celebrity effect on live-streaming e-commerce and analyses its strengths, weaknesses, opportunities and threats through SWOT analysis with typical case studies in China. The findings highlight the key role of celebrity endorsement in increasing consumer engagement and purchase intention, while also exploring the potential risks and challenges. The paper concludes with suggestions for optimisation and development to maximise the role of the celebrity effects in live-streaming e-commerce.

Point of Interest Recognition and Tracking in Aerial Video during Live Cycling Broadcasts

Road cycling races, such as the Tour de France, captivate millions of viewers globally, combining competitive sportsmanship with the promotion of regional landmarks. Traditionally, points of interest (POIs) are highlighted during broadcasts using manually created static overlays, a process that is both outdated and labor-intensive. This paper presents a novel, fully automated methodology for detecting and tracking POIs in live helicopter video streams, aiming to streamline the visualization workflow and enhance viewer engagement. Our approach integrates a saliency and Segment Anything-based technique to propose potential POI regions, which are then recognized using a keypoint matching method that requires only a few reference images. This system supports both automatic and semi-automatic operations, allowing video editors to intervene when necessary, thereby balancing automation with manual control. The proposed pipeline demonstrated high effectiveness, achieving over 75% precision and recall in POI detection, and offers two tracking solutions: a traditional MedianFlow tracker and an advanced SAM 2 tracker. While the former provides speed and simplicity, the latter delivers superior segmentation tracking, albeit with higher computational demands. Our findings suggest that this methodology significantly reduces manual workload and opens new possibilities for interactive visualizations, enhancing the live viewing experience of cycling races.

Identification and location monitoring through Live video Streaming by using blockchain

Identification and location monitoring through Live video Streaming by using blockchain

A comprehensive investigation of an LTE-enabled smart door system using the Arduino UNO

Traditional residential security systems frequently lack remote access and real-time reaction capabilities, necessitating home automation security technology developments. With the rise of the Internet of Things (IoT), solutions such as the Smart Door are prepared to transcend these constraints by leveraging modern communication technology. The article aims to develop, implement, and evaluate an LTE-enabled Smart Door system built on the Arduino Uno platform. The goal is to improve home security by offering keyless entry, real-time surveillance, and remote operation, increasing user convenience and system reliability. We created a Smart Door system with an Arduino Uno microcontroller and LTE connectivity. The system's performance was measured using a variety of measures, including response time, reliability, and security against unwanted access. Testing included simulation environments and real-world scenarios to ensure the system's operational efficacy and security. The Smart Door system performed admirably, with a reaction time of less than two seconds in 95% of commands and strong resilience to brute force attacks. It effectively incorporated keyless entry, live video streaming, and real-time alarms, all controlled by a smartphone app, demonstrating the system's dependability and efficiency in live scenarios. The LTE-enabled Smart Door system considerably improves over traditional home security options. It provides enhanced security, simple operation, and more outstanding remote management capabilities by utilizing IoT technology. These enhancements fill essential holes in existing systems, providing a safer, dependable, and convenient answer for modern home security requirements.

Usability, Ergonomics, and Educational Value of a Novel Telestration Tool for Surgical Coaching: Usability Study.

Telementoring studies found technical challenges in achieving accurate and stable annotations during live surgery using commercially available telestration software intraoperatively. To address the gap, a wireless handheld telestration device was developed to facilitate dynamic user interaction with live video streams. This study aims to find the perceived usability, ergonomics, and educational value of a first-generation handheld wireless telestration platform. A prototype was developed with four core hand-held functions: (1) free-hand annotation, (2) cursor navigation, (3) overlay and manipulation (rotation) of ghost (avatar) instrumentation, and (4) hand-held video feed navigation on a remote monitor. This device uses a proprietary augmented reality platform. Surgeons and trainees were invited to test the core functions of the platform by performing standardized tasks. Usability and ergonomics were evaluated with a validated system usability scale and a 5-point Likert scale survey, which also evaluated the perceived educational value of the device. In total, 10 people (9 surgeons and 1 senior resident; 5 male and 5 female) participated. Participants strongly agreed or agreed (SA/A) that it was easy to perform annotations (SA/A 9, 90% and neutral 0, 0%), video feed navigation (SA/A 8, 80% and neutral 1, 10%), and manipulation of ghost (avatar) instruments on the monitor (SA/A 6, 60% and neutral 3, 30%). Regarding ergonomics, 40% (4) of participants agreed or strongly agreed (neutral 4, 40%) that the device was physically comfortable to use and hold. These results are consistent with open-ended comments on the device's size and weight. The average system usability scale was 70 (SD 12.5; median 75, IQR 63-84) indicating an above average usability score. Participants responded favorably to the device's perceived educational value, particularly for postoperative coaching (agree 6, 60%, strongly agree 4, 40%). This study presents the preliminary usability results of a novel first-generation telestration tool customized for use in surgical coaching. Favorable usability and perceived educational value were reported. Future iterations of the device should focus on incorporating user feedback and additional studies should be conducted to evaluate its effectiveness for improving surgical education. Ultimately, such tools can be incorporated into pedagogical models of surgical coaching to optimize feedback and training.

Reform of Photography Education in Chinese Applied Universities in the Self-Media Era: Evidence from Responsible Persons in the Photography Major

Live streaming and short videos have become the most favored means of communication, and the prevalence of self-media has had a profound impact on traditional media industries, indirectly promoting reforms in Chinese photography education. This study delves into the current state of photography education in Chinese applied universities and the transformations undertaken in response to the challenges posed by self-media. The research covers the development status of university photography programs, the impact of live streaming and short videos on photography education, changes in teaching models, and the issues encountered during these reforms. Through in-depth interviews with heads and instructors of photography programs at five applied universities in China, and using NVIVO software for thematic analysis, the study examines the professional divisions, curriculum design, teaching methods, student employment rates, reform policies, and the challenges faced. The research finds that the prevalence of live streaming and short videos has had a positive impact on the dynamic imaging specialties within photography education, but has been detrimental to the static imaging specialties. Universities have actively reformed their teaching methods in response to market demands, yet still face challenges in integrating live streaming-related courses. This study provides the latest insights for practitioners in the Chinese self-media industry and offers valuable references for university policymakers and researchers.

Digital Twin-Based Alternate Ego Modeling and Simulation: Eva Herzigová as a 3D MetaHuman Avatar

Abstract Due to Dimension Studios and Unsigned Group, by likeness and movement digitization, the human supermodel Eva Herzigová’s highly realistic metahuman and 3D digital twin lead to perpetual career extension (e.g., on virtual runway shows by catwalk replication or in advertising campaigns by particular pose adoption), including in virtual worlds such as metaverse spaces, by use of extended reality technologies, in terms of virtual clothing, digital hair, skin texture, facial and body expressions, and makeup styling as avatar appearance and persona. In this paper, we show how, by likeness capturing (e.g., signature walk by motion capture shoot for face and body movement tracking) and virtual clothing, Eva Herzigová’s hyperreal 3D avatar (lifelike 3D digital human clone or realistic virtual human) is configured with subsequent restyled hair and make-up, 3D clothing modeling, and extended reality fashion shows on virtual catwalks or campaign shoots. We clarify that photographic reference materials and a sensor-based motion capture suit assisted Metahuman Creator in capturing facial and motion data for Eva Herzigová’s digital avatar, configuring realistic depiction and virtual activations of beauty standards, in addition to digital likeness, signature style, talent representation, and narrations, typifying digital human creation and presence in virtual worlds (e.g., for metaverse-based shoots). We conclude that Eva Herzigová’s hyperrealistic 3D digital human twin and MetaHuman avatar operate as a styled virtual alter ego and versatile personal brand extension by motion capture technologies for immersive virtual fashion shows, live video streams and broadcast, and digital wearable modeling.

A Lightweight Visual Understanding System for Enhanced Assistance to the Visually Impaired Using an Embedded Platform

Visually impaired individuals often face significant challenges in navigating their environments due to limited access to visual information. To address this issue, a portable, cost-effective assistive tool is proposed to operate on a low-power embedded system such as the Jetson Nano. The novelty of this research lies in developing an efficient, lightweight video captioning model within constrained resources to ensure its compatibility with embedded platforms. This research aims to enhance the autonomy and accessibility of visually impaired people by providing audio descriptions of their surroundings through the processing of live-streaming videos. The proposed system utilizes two distinct lightweight deep learning modules: an object detection module based on the state-of-the-art YOLOv7 model, and a video captioning module that utilizes both the Video Swin Transformer and 2D-CNN for feature extraction, along with the Transformer network for caption generation. The goal of the object detection module is for providing real-time multiple object identification in the surrounding environment of the blind while the video captioning module is to provide detailed descriptions of the entire visual scenes and activities including objects, actions, and relationships between them. The user interacts via a headphone with the proposed system using a specific audio command to trigger the corresponding module even object detection or video captioning and receiving an audio description output for the visual contents. The system demonstrates satisfactory results, achieving inference speeds between 0.11 to 1.1 seconds for object detection and 0.91 to 1.85 seconds for video captioning, evaluated through both quantitative metrics and subjective assessments.

Video stalling identification for web live streaming under HTTP-FLV

Live broadcasts have become one of the most popular forms of entertainment. Quality of user Experience (QoE) is a vital quantitative criterion for evaluating user satisfaction while watching live broadcasts, and it is positively correlated with the increase in the income of Internet Service Providers (ISPs). Video stalling identification plays a crucial role in the evaluation of QoE. However, encrypted live streaming hides video content, which makes identifying video stalling challenging. Existing studies primarily detect video stalling in a fixed time interval and focus on high-dimensional features. However, the capacity of the client byte buffer is dynamic, resulting in the stalling and non-stalling existing in a certain and fixed stalling time. In addition, the handling time of abundant features causes further latency. We propose Truncation of Dynamic Bytes and non-linear Integrated Modification based on Double Buffers (DB2) to identify video stalling under HTTP-FLV protocol in various network conditions and live types. We pull real-time video to get client buffer parameters and build a dynamic mapping based on the double buffer between network packets and the video playing states. This allows a more objective and precise evaluation of video stalling. We design a new network feature by creating a non-linear relationship between network packets and the client buffer. This is achieved by combining non-linear convergent distribution with basic traffic features. The feature is fed into a lightweight machine learning model to train the classifier, achieving low processing latency and high identification accuracy. The experimental results show that DB2 can achieve 98.91% stalling identification accuracy with 1.256 ms operation time in a mixture of live video types, outperforming state-of-the-art techniques.

Leveraging Edge Computing for Video Data Streaming in UAV-Based Emergency Response Systems.

The rapid advancement of technology has greatly expanded the capabilities of unmanned aerial vehicles (UAVs) in wireless communication and edge computing domains. The primary objective of UAVs is the seamless transfer of video data streams to emergency responders. However, live video data streaming is inherently latency dependent, wherein the value of the video frames diminishes with any delay in the stream. This becomes particularly critical during emergencies, where live video streaming provides vital information about the current conditions. Edge computing seeks to address this latency issue in live video streaming by bringing computing resources closer to users. Nonetheless, the mobile nature of UAVs necessitates additional trajectory supervision alongside the management of computation and networking resources. Consequently, efficient system optimization is required to maximize the overall effectiveness of the collaborative system with limited UAV resources. This study explores a scenario where multiple UAVs collaborate with end users and edge servers to establish an emergency response system. The proposed idea takes a comprehensive approach by considering the entire emergency response system from the incident site to video distribution at the user level. It includes an adaptive resource management strategy, leveraging deep reinforcement learning by simultaneously addressing video streaming latency, UAV and user mobility factors, and varied bandwidth resources.

The influence of live video hosts' speech act on purchase behaviour

With the broad adoption of internet, China's e-commerce industry has grown vigorously. The influence mechanism between the host's speech act and the consumer's purchase behaviour deserves further exploration. Based on speech act theory, this study divided the speech acts into four categories: representative, expressive, commissive and directive. We established a model investing the effect of different speech acts on customer engagement and purchase behaviour. We recorded live streaming video and transferred the voice to the text. Our research shows that the representative act and the expressive act negatively affect purchase behaviour, whereas the commissive act has a positive impact on purchase behaviour. Furthermore, customer engagement mediates the negative impact of the directive act and the representative act on purchase behaviour. The conclusions can provide guidance for enhancing the speech skills of the hosts in live streaming.