State-of-the-Art Applications in Computer Vision for Real-Time Target Detection: A Comprehensive Overview

State-of-the-Art Applications in Computer Vision for Real-Time Target Detection: A Comprehensive Overview

Prof.Ms. Rashmi Sonar1 , Prof. Vinay Rajgure2, Pranali B. Tobre3

Scholars, Department of Computer Science and Engineering3

Assistant Professor, Department of Computer Science and Engineering1,2

Prof Ram Meghe College Of Engineering & Management, Badnera, Amravati, Maharashtra, India rashmi.sonar@prmceam.ac.in1, vinay.rajgure@gmail.com2, pranali.tobre@gmail.com3

Abstract: This study presents a computer vision module for real-time target detection in live video feeds utilising a deep learning object detector based on the YOLO technique. By pre-training on a sizable target picture dataset, adding a lightweight attention mechanism, and using post-processing to eliminate false positives and improve bounding boxes, it improves YOLO's accuracy. Modern precision is attained by the module on difficult real-world video[4]. It also examines new developments in real-time object identification, such as transfer learning, noise and occlusion resistance, and innovative structures. This programme has potential applications in robotics, surveillance, and augmented reality.

Keywords: Computer vision, real-time target recognition, YOLO algorithm, deep learning, object detection, accuracy improvement, live camera, real-world footage.