Real-Time Object Detection – YOLO Computer Vision

About the project

Developed a real-time object detection module during an AI & Machine Learning internship at ROBX AI, designed to identify and locate multiple objects directly from a live video stream. The system combines YOLO-based object detection with an OpenCV video-processing pipeline to perform inference continuously while maintaining the responsiveness required for real-time applications.OpenCV handles frame capture, preprocessing, and visualization of detected objects, while YOLO performs single-pass inference to generate object classes, confidence values, and bounding-box locations. The detection stage was explored using both TensorFlow and PyTorch implementations to better understand differences in model execution and the trade-off between detection accuracy and inference speed.The result was a functional live detection pipeline and practical experience optimizing computer-vision models for real-time rather than offline video processing.