Al-time object detection and tracking app using YOLOv8 and the SORT algorithm. It detects objects in images, videos, or webcam feeds, draws bounding boxes with class/ID, and tracks them across frames for consistent identification and visual output.

A system that detects scene text on traffic signs through images and videos

Crosswalk & Traffic Sign Recognition Dataset: A labeled collection of 1,800 images for autonomous driving. Features 14 classes including directional signs, obstacles, and road conditions. Initial models achieved a 78.1% mAP.

Utilizing synthetic images to enhance the automated recognition of small-sized construction tools

REST API to detect objects in images. Get labels, confidence scores, and bounding box coordinates. Powered by neural networks.

YOLOv8-based detection & automated counting of palm oil fruit bunches (TBS) — classifying Matang, Mentah, and Cacat in real-time. Industry collaboration project with PTPN, developed at Binus University. Exported to ONNX for edge deployment.

A new benchmark model for the automated recognition of a wide range of heavy construction equipment