FIX-R Web Platform
Real-Time Computer Vision Interface
An intelligent diagnostic platform that uses machine learning and augmented reality to identify and troubleshoot issues in cars, boats, and home systems. It offers hierarchical issue diagnosis, repair instructions, product recommendations, downloadable PDF guides, and AR-guided repairs. The platform includes user authentication, admin controls, continuous learning capabilities, and diagnostic history tracking.
Developed a web-based interface for real-time object detection and AI-assisted diagnostics.
-
Implemented real-time object detection (COCO-SSD via TensorFlow.js) directly in the browser
-
Built a camera pipeline with bounding boxes and confidence overlays for interactive user feedback
-
Designed a validation system comparing real-time detections with ML model outputs to improve accuracy
-
Optimized performance using frame throttling (30fps) and efficient client-side processing
-
Integrated frontend with backend APIs for diagnostics, history tracking, and repair workflows