This paper presents a real-time object detection system designed to assist visually impaired individuals through audio alerts. The system is implemented as a web-based application using HTML, CSS, and JavaScript, making it accessible and easy to use. It utilizes TensorFlow.js along with the COCO-SSD pre-trained model to detect objects from live video captured through a camera. The system identifies common objects such as people, vehicles, and everyday items, and determines their position relative to the user. Based on the detection results, appropriate alert messages are generated and converted into speech using the Web Speech API. This enables users to receive real-time information about their surroundings without relying on visual input. The system operates continuously and provides quick responses with minimal delay. Overall, the proposed solution is cost-effective, efficient, and demonstrates the practical application of machine learning in assistive technology. The system is designed to function in real-world environments and can handle multiple object detections simultaneously. Its web-based nature ensures portability and ease of access across devices. This approach highlights the potential of integrating modern web technologies with artificial intelligence for developing practical assistive solutions. The system supports real-time processing and can detect multiple objects simultaneously. It is designed to be user-friendly and accessible across different devices through a web browser. This approach demonstrates how machine learning and web technologies can be effectively combined to create a practical assistive solution.

To evaluate the efficacy of a teledentistry-based follow-up approach compared with standard in-person follow-up for monitoring oral health among patients with head and neck cancer (HNC) after the completion of radiotherapy, using a noninferiority framework. In this two-arm, noninferiority randomized controlled trial, teledentistry was compared with standard in-person follow-up in patients with HNC after the completion of radiotherapy. The participants were randomized 1:1 and followed up at 1, 3, and 6 months. For teledentistry, a secure digital platform incorporating live video and self-photography was used. The primary outcomes were the percentage of tooth surfaces with plaque (plaque index) and bleeding on probing. The secondary outcomes were radiotherapy-related oral complications (i.e., dental caries incidence, trismus assessed via maximal mouth opening, mucositis severity, xerostomia severity, and osteoradionecrosis incidence) and patient-reported measures, including oral health-related quality of life (Oral Health Impact Profile, OHIP-14) and patient satisfaction. Noninferiority was assessed using analysis of covariance and risk difference calculations, comparing outcomes between groups based on data collected 2 weeks after the 6-month follow-up. A predefined noninferiority margin of 15%, representing the maximum clinically acceptable difference between groups, was applied with 95% confidence intervals (CI). Twenty-seven participants completed the study (teledentistry group: n = 14; standard group: n = 13), with no statistically significant baseline differences. Teledentistry was noninferior to standard follow-up for the plaque index (mean difference: 1.9% points; 95% CI [- 8.6, 12.4]) and bleeding on probing (mean difference: -2.2% points; 95% CI [- 17.2, 12.8]). The incidence of osteoradionecrosis was also noninferior (risk difference: -0.15; 95% CI [- 0.35, 0.04]). Patient satisfaction was high across the groups. Outcomes related to other radiotherapy-related oral complications and oral health-related quality of life were inconclusive. Teledentistry appears to be a feasible and noninferior follow-up modality for patients with HNC during the initial 6 months after radiotherapy. TCTR20221008001, registered on October 8, 2022.

Autonomous driving offers a promising way to tackle the rising number of fatalities from traffic accidents. An autonomous vehicle includes many features, but the ability to detect pedestrians is crucial, challenging, and relevant to various real-time situations like surveillance, tracking people, and monitoring. Accurately identifying pedestrians is difficult because they can appear in different shapes, positions, and postures. They can wear various types of clothing and sometimes be partially hidden or blend in with nearby objects. This paper focuses on the real-time detection of pedestrians for self-driving cars using a popular hardware platform: The field programmable gate array (FPGA), Ultra 96 v2. The study implements a method for pedestrian detection based on a histogram of oriented gradients (HOG) combined with a support vector machine (SVM) classifier to recognize individuals on the FPGA board, leveraging high-level synthesis (HLS) tools. The effectiveness of the system has been tested on both still images and live video. The results show that advanced FPGA boards like the Ultra 96 v2 significantly improve performance metrics. The system operates at a clock frequency of 150 MHz while using less than half of the available resources and consuming around 2.5 W of power. Also, the system reports the pedestrian detection accuracy close to 95% and other efficient metrics for detection evaluation, like precision (78.6%), recall (88.3%), and F1 Score (83.1%). In summary, the developed system can detect pedestrians in real-time and has the potential to significantly improve the development of a smart and safe transportation environment.

The rapid evolution of human-computer interaction (HCI) technologies has created demand for natural, contactless interfaces that transcend the limitations of traditional input devices. This paper presents the design and implementation of a real-time gesture-based interaction system that integrates Google’s MediaPipe Hands framework with the Three.js WebGL rendering library. The proposed system, titled Three.js Hand Recognition Panel, captures live video from a standard webcam and detects up to two hands simultaneously, extracting 21 three-dimensional landmark points per hand. Distancebased and positional gesture recognition algorithms map these landmarks to interactive commands that drive a WebGL-rendered 3D virtual environment. Operating entirely within a web browser without specialized hardware, the system achieves 50–60 FPS on standard desktops, gesture recognition accuracy exceeding 88%, and average response latency below 70 ms. Experimental results validate the feasibility of deploying production-quality gesture interfaces through open-source web technologies. The work contributes a modular, extensible architecture suitable for applications in gaming, virtual reality, healthcare, and assistive technologies.

In recent years, the demand for remote communication and online collaboration solutions has increased significantly, particularly in the fields of education, recruitment, and professional assessment. Organizations and educational institutions are rapidly shifting toward digital platforms to conduct interviews, technical assessments, and candidate evaluations efficiently. Traditional interview methods often involve several challenges such as geographical limitations, travel costs, scheduling conflicts, infrastructure dependency, and timeconsuming coordination processes. These issues become even more difficult when organizations need to conduct interviews with candidates located in different cities or countries. To overcome these limitations and support modern recruitment requirements, the TalentIQ Interview Management System has been developed as a smart and efficient web-based interview platform. The TalentIQ platform is designed to facilitate real-time online technical interviews directly through a web browser without requiring any external or third-party video conferencing software. The system enables interviewers and candidates to communicate seamlessly through live video calls, instant messaging, and collaborative coding environments within a single integrated platform. By providing all interview-related functionalities in one centralized system, the platform simplifies the overall interview process and improves user experience for both interviewers and candidates.

To construct a reliable robot-assisted radical prostatectomy (RARP) surgical phase-recognition model and sought to deploy the model within the SurgSmart platform (Chengdu Withai Innovations Technology Co., Ltd., Chengdu, China) to explore its feasibility and preliminary clinical utility in real surgical workflows, including its usability, interpretability, and potential value for surgical education and quality review. A total of 72 complete RARP procedure videos were collected, and divided into training, verification and test groups (7:1:2). We adopted Hiera (Meta Platforms, Inc., Menlo Park, CA, USA), a hierarchical vision transformer, as the backbone model for surgical phase recognition. Model performance was evaluated using precision, recall, F1-score, and overall accuracy. The trained model was deployed on the SurgSmart platform and tested in two real world RARP procedures to evaluate its feasibility for intraoperative and postoperative use. Temporal annotation quality was high, with a mean inter-annotator Intersection over Union (IoU) score of 0.99 and a weighted IoU score of 0.97. Based on the finalised annotations, the trained Hiera model achieved a weighted F1-score of 0.91, a macro F1-score of 0.90, and an overall accuracy of 0.91 on the test set. Across the seven surgical phases, 'urethral anastomosis' and 'intrafascial dissection' reached F1-scores of 0.96 and 0.92, and the remaining phases demonstrated F1-scores within the 0.85-0.96 range. The confusion matrix demonstrated that most surgical phases were correctly classified, with a high concentration of samples along the diagonal cells, indicating strong alignment between predicted and ground-truth labels. During real-time deployment, the system processed live surgical video streams continuously and generated phase predictions throughout the operation without interruption. All outputs were produced without processing errors or interface interruptions, confirming stable operation in both real-time and retrospective modes. Accurate surgical phase recognition might be achievable in RARP under a limited but high-quality data setting, although further validation with larger and more diverse datasets is needed to confirm these findings.

The proposed Smart Surveillance System presents a novel, hardware-integrated prototype demonstration aimed The proposed Smart Surveillance System presents a groundbreaking hardware-integrated prototype that decisively validates the effectiveness of a dual-branch anti-spoofing model on the low-power edge device, Raspberry Pi 3B+. This prototype goes beyond algorithmic performance research, showcasing a fully functional proof of concept. In contrast to existing surveillance solutions that typically rely on centralized cloud processing or basic recognition systems, our system employs an advanced dual-branch model that utilizes both spatial and frequency-domain features. This approach enables real-time anti-spoofing with an impressive error rate of less than 2%. What truly sets our system apart is its seamless end-to-end integration of cloud-based authentication, edge-level inference for rapid response, and an interactive live video conferencing feature. This configuration empowers immediate verification and action during potential spoofing events. With IoT-enabled devices, our system ensures effortless communication for live streaming, automated alerts, and scalable cloud data management. Coupled with edge computing, it guarantees real-time decision-making with minimal latency. Experimental results confirm its high accuracy in distinguishing genuine users from spoofing attempts, positioning our solution as a lightweight, proactive, and user-interactive surveillance option that is perfectly suited for homes, enterprises, and public infrastructures.

ABSTRACT This paper presents the design and implementation of BUBBLE, a low-cost, small-scale remotely operated underwater vehicle (ROV) for shallow-water observation and subsea inspection in resource-limited maritime environments. The system integates a lightweight 3D-printed hull reinforced with fiberglass, an STM32-based embedded control architecture, and a Python-based supervision interface enabling live video streaming and real-time sensor monitoring over a tethered link. Experimental validation through controlled basin tests demonstrates watertightness, propulsion feasibility, and stable real-time monitoring.The total fabrication cost is approximately USD 520, making BUBBLE accessible for education, environmental monitoring, and applied research where conventional ROVs remain financially or logistically inaccessible. This work contributes to cost-efficient subsea engineering by demonstrating a scalable, locally manufacturable alternative to commercial ROVs. Selected footage is available at: https://youtu.be/V3rBI19AnGs.

Enhancing Cataract Surgery Training Through a Secure Microscope-Integrated Live Video Supervision System.

PCCUA: An attention-based prediction-driven joint collaborative caching and user association algorithm for live video streaming in edge networks

Gut-directed hypnotherapy is an effective treatment for patients with irritable bowel syndrome (IBS). Group delivery and nurse-led hypnotherapy can increase availability. Online treatment shows promising results, but this has not been tested in a group format. To investigate the acceptability and efficacy of nurse-led, online group hypnotherapy in patients with IBS. Patients received eight sessions of gut-directed hypnotherapy in groups via live video conferencing. IBS symptoms were assessed at baseline, mid-treatment, after treatment, and at follow-up. Patients who reported an IBS-SSS reduction of ≥ 50 points were considered responders. Extracolonic symptoms, psychological symptoms, and quality of life were assessed, as well as usability and treatment satisfaction. The study results were compared to previous assessments of group hypnotherapy delivered on-site. After hypnotherapy, patients were asked which treatment modality (online or on-site) they would prefer. We included 51 patients. IBS severity was reduced after hypnotherapy (median IBS-SSS: 304 (225-385) vs. 225 (172-312), p < 0.001), and 27 patients (53%) were responders. These results are comparable to on-site group hypnotherapy outcomes; IBS-SSS: 310 (232-368) versus 230 (151-330), p < 0.001, responders: 55%. Symptom reduction was sustained at six-month follow-up. Quality of life, extracolonic, and psychological symptoms also improved. The patient ratings of the usability of the video call platform and treatment satisfaction were high. Nurse-led, gut-directed group hypnotherapy delivered online is acceptable, often preferred by patients, and has comparable efficacy to in-person group hypnotherapy. By combining group and online treatment, hypnotherapy can be made more accessible for patients.

Falls from height remain a critical safety challenge in the construction industry, where vision-based monitoring systems must balance accuracy, efficiency, and reliability. This study proposes and validates an end-to-end framework for proactive fall prevention that directly addresses this trade-off. First, the framework introduces an efficiency-aware detector, You Only Look Once version 8s for Construction Worker Safety (YOLOv8s-CWS), with targeted architectural enhancements to improve accuracy on small and occluded workers. Second, it establishes a synergistic design where this high-precision detector is paired with the motion-centric ByteTrack algorithm to achieve reliable identity tracking amidst the visual clutter of dynamic construction sites. Finally, it incorporates a validated, building-information-modeling-agnostic method for real-time hazard judgment, enabling safety personnel to define unprotected edges and openings directly on a live video feed. System performance was validated on challenging construction site video sequences. The proposed YOLOv8s-CWS detector achieved a mean average precision of 83.7%, a +1.5% point improvement over its baseline, while operating at 142.7 frames per second. This enhanced detection directly mitigated tracking failures, boosting the system’s identity F1 score to 75.32%, for a +7.85 point gain over the baseline configuration. Field deployments confirmed the framework’s effectiveness in delivering robust, real-time intrusion alerts. This study provides a computationally efficient and validated solution for proactive safety monitoring, offering a practical tool to mitigate fall-related risks on construction sites.

This project presents an AI-based real-time fitness assistance system that provides personalized workout guidance using computer vision techniques. The system uses a standard webcam to capture live exercise videos, eliminating the need for wearable sensors or specialized hardware. Human pose estimation is performed using MediaPipe Pose to accurately detect and track body landmarks in real time. A React-based web application is used to manage user interaction, workout selection, and live visualization of exercise performance. The detected landmarks are connected to form a skeletal representation of the human body, which enables detailed posture and movement analysis. Joint angles are calculated from the skeletal model to evaluate exercise correctness and body alignment. Each performed exercise is compared with predefined reference poses to assess posture accuracy. The system computes an accuracy score based on deviations between the user’s posture and the ideal pose. Real-time visual and textual feedback is provided to help users correct improper movements during workouts. The system continuously monitors workout duration and movement intensity throughout the session. Calorie expenditure is estimated using a metabolic equivalent–based model combined with motion analysis. Posture accuracy and movement quality are used to refine the calorie estimation for improved realism. User performance metrics are recorded at the end of each session. Historical workout data is stored to support long-term progress tracking and performance evaluation. Analytical visualizations are used to present trends in accuracy, duration, and calorie burn. All pose detection and analysis are performed on the client side within the web browser. This client-side approach ensures low latency and preserves user privacy by avoiding video data transmission. The system supports multiple workout types such as yoga, gym exercises, and dance-based fitness routines. The proposed solution is scalable and accessible across devices with a web camera. Overall, the system demonstrates an effective and intelligent approach to AIdriven home-based fitness training.

Exploring unknown or dangerous environments without risking human life is becoming increasingly important, and this project focuses on developing a smart and affordable AI-powered rover capable of navigating outdoor terrains while detecting obstacles in real time. The rover uses a Raspberry Pi 5 as its brain, along with a camera and sensors that help it understand its surroundings. A machine learning model is used to recognize obstacles, while an ultrasonic sensor measures distance and an IMU maintains balance and stability. By combining all this information, the rover can make quick decisions, avoid collisions, and adjust its movement based on terrain conditions such as grass, soil, and gravel. A key feature of the system is its flexibility, as it can operate autonomously or be controlled manually through a simple web-based interface that provides live video streaming. This allows human intervention whenever required. The rover was tested in real outdoor conditions, where it demonstrated smooth movement, reliable obstacle detection, and adaptability to different surfaces. Overall, the project shows that an intelligent and effective autonomous system can be built using low-cost components, making it suitable for applications like surveillance, exploration, and research while reducing risks to human life.

This paper presents the implementation of a Hexacopter Unmanned Aerial Vehicle (UAV) designed for real-time surveillance, load carrying, and environmental monitoring applications. The hexacopter configuration offers improved stability, higher payload capacity, and better maneuverability compared to conventional quadcopters. The proposed system integrates a flight controller, GPS module, camera system, sensors, and communication modules to perform multiple tasks efficiently. For surveillance, the UAV provides live video streaming and remote monitoring of targeted areas. For load carrying, it is capable of transporting small payloads such as medical supplies, packages, or emergency materials. In environmental monitoring, sensors are used to measure parameters such as temperature, humidity, gas concentration, and air quality in real time. The system is tested under different operating conditions to evaluate flight stability, payload performance, and data accuracy. Experimental results show that the hexacopter UAV is reliable, cost-effective, and suitable for applications in disaster management, agriculture, security, and smart city operations. The project demonstrates the potential of multifunctional UAV systems for modern real-time monitoring transportation needs

Conventional approaches for tracking devices, including manual logs, periodic inspections, and RFID-based monitoring, are still widely used in many organizational environments. However, these methods suffer from several limitations such as dependency on human effort, risk of data inaccuracies, and lack of effective mechanisms to prevent misuse or unauthorized handling. In large-scale deployments, these systems fail to provide continuous visibility of device location and operational status. To overcome these challenges, this work presents a real-time device tracking framework that continuously monitors device movement using sensor-driven data acquisition. The system utilizes technologies such as GPS modules, IoT-based sensing, and wireless communication to capture and process tracking data without requiring active user involvement. This enables accurate, real-time monitoring and improves overall asset visibility and control. The proposed solution is implemented entirely in Python and integrates a lightweight database with an interactive web-based dashboard for visualization and management. The system also incorporates anomaly detection to identify irregular device behavior and enhance security. Experimental evaluation demonstrates high tracking accuracy, low latency, and a significant reduction in administrative workload. The framework is scalable, cost-effective, and suitable for deployment across academic, industrial, and enterprise environments. Methods: The system is built exclusively in Python and integrates multiple technologies to enable continuous, real-time device tracking and monitoring. It utilizes IoT-based data acquisition (such as GPS modules or network signals) along with OpenCV for optional visual tracking and live video streaming where applicable. Device identification is achieved through unique identifiers and feature-based representations, enabling accurate tracking across environments. Additionally, anomaly detection mechanisms are incorporated to verify device authenticity and detect irregular movement patterns, enhancing system reliability and security.

Rapid urbanization has led to a significant increase in traffic congestion and noise pollution in metropolitan areas. Excessive honking and traffic signal violations are common issues at busy intersections, contributing to unsafe road conditions and environmental noise. This paper presents Hush-Traffic, an IoT-based smart traffic monitoring system designed to detect excessive honking and monitor traffic rule violations in real time. The proposed system integrates a sound sensor, an Arduinobased traffic controller, and an ESP32-CAM module to create an intelligent monitoring platform. The sound sensor continuously measures environmental noise levels near traffic signals, and when the detected noise exceeds a predefined threshold, the system identifies it as a potential violation. The microcontroller then triggers the ESP32-CAM module to capture images and stream live video through a localhost-based web interface. The traffic signal operates with predefined timing sequences while the monitoring system records abnormal events for analysis. A web-based dashboard provides real-time visualization of signal status and camera output. Experimental results demonstrate that the proposed system effectively detects excessive honking and enables efficient monitoring of traffic behavior. The system offers a low-cost, scalable, and practical solution for improving traffic discipline and reducing noise pollution in smart city environments.

This project introduces an automated Examination Malpractice Detection System that utilizes YOLOv8 for real-time object detection and MediaPipe for human pose estimation. By processing live video streams through a Flask-based backend, the system identifies prohibited items and suspicious behavioral patterns such as excessive head movement or unauthorized communication. To ensure high reliability, it employs a 60% confidence threshold and a decision-fusion layer that reduces false positives by analyzing sequences of motion. The architecture features a seamless administrative response layer that captures time-stamped evidence and compiles infractions into a secure zip archive. Upon session completion, the system automatically transmits a comprehensive summary and evidence file to examiners via an integrated email notification service, ensuring a transparent and objective disciplinary review process.

This study aims to examine the forms and functions of commissive speech acts in political discourse found in live broadcast videos of the Red and White Cabinet Meeting uploaded on YouTube. Using a qualitative descriptive approach, the research focuses on utterances containing commissive speech acts delivered by meeting participants during the sessions. Data collection was conducted through observation and note-taking, while data analysis applied Charles Morris’s pragmalinguistic model, including data collection, categorization, analysis, and conclusion drawing. The results reveal 20 instances of commissive speech acts, consisting of promises (45%), intentions (25%), offers (10%), and guarantees (20%). Promises and intentions are the most dominant, functioning to provide certainty, demonstrate commitment, reassure, and build public trust. Offer speech acts are used to propose policies, whereas guarantees aim to convince interlocutors and the public about the continuity of government programs. These findings indicate that commissive speech acts play a strategic role in political communication to strengthen legitimacy, commitment, and trust. The study further suggests that the use of commissive speech acts in digital spaces serves as an instrument for public accountability.

The Intelli Spy Robot with Live Streaming and Location Surveillance is an advanced autonomous system designed to perform real-time environmental monitoring, hazard detection, and remote surveillance in areas that are inaccessible or dangerous for human personnel. The proposed system integrates an ESP-32 microcontroller with multiple sensing units including a fire sensor, gas sensor, and metal sensor, alongside an ESP-CAM module for high-definition live video streaming, and a GPS module for precise location tracking. The robot is powered by a dedicated rechargeable battery unit (RPS) ensuring uninterrupted field operation. Upon detection of hazardous conditions such as fire outbreaks, toxic gas leakage, or concealed metallic objects, the system instantly triggers audio alerts via a buzzer and displays critical data on an onboard LCD screen. All processed information is simultaneously transmitted to a centralized IoT platform, enabling remote operators to monitor, analyze, and respond to threats in real time from any geographic location. The system eliminates human risk in hostile environments by providing continuous, automated, and intelligent surveillance. It is designed to be deployable in military zones, disaster-affected areas, industrial sites, and border patrol regions. The integration of robotics, embedded intelligence, and cloud connectivity makes this system a comprehensive solution for modern surveillance challenges. This project represents a significant step forward in combining embedded systems, computer vision, wireless communication, and sensor fusion for autonomous field operations.