AI VISION INSPECTION TO ELIMINATE MANUAL CHECKS 

Computer Vision-Based Inspection System for Product Verification is a Capstone project developed by Team PassME from RMIT University in collaboration with industry partner P&G.

The project delivers a responsive web prototype integrated with a YOLOv8 deep learning model to automate defect detection and streamline quality control. The system automatically detects and localizes visual defects, providing annotated images with bounding boxes, labels, confidence scores, and critical PASS/FAIL or TAMU-related decision support.



Key features of the web platform include image/dataset upload, interactive browser-based bounding-box annotation, defect class management, administrator-controlled model training, and YOLO-compatible dataset export. Designed to optimize production inspection, this system minimizes manual verification, ensures evaluation reproducibility, and effectively addresses industrial data scarcity challenges