Automated exam analysis and cheating detection system. Uses OCR & ArUco for extraction and DBSCAN + PCA to flag suspicious student patterns.
Open-source Linux gallery app. Features a trained MobileNet classifier for auto-categorization and Levenshtein distance for typo-tolerant search.
Real-time perception module. Deployed YOLOv5 on Raspberry Pi for object detection and implemented an OpenCV pipeline for lane segmentation.