I’m developing a focused “Google Lens-style” feature that takes a photo (or live camera feed), recognises everyday objects and consumer products, and then immediately returns detailed information about each recognised item. The scope is limited to pure image recognition and analysis—no text extraction, translation, or AR overlays at this stage. You’ll design, train, and integrate the object-detection model, then connect it to a lightweight information layer that surfaces concise descriptions, specifications, and any metadata I provide. Whether you prefer Google Cloud Vision, TensorFlow, PyTorch, OpenCV, or another modern framework, the end result must run quickly and accurately on a mobile device or web backend. Deliverables I need to see: • A working prototype (Android, iOS, or web) that draws bounding boxes around detected objects/products and shows the corresponding info panel • Model files, inference code, and clear documentation so I can retrain with additional categories later • Setup guide plus a short report detailing achieved accuracy and recommended next steps I’ll measure success by recognition speed, precision (aiming for 90 %+ top-1 accuracy on my test set), and code clarity. Let me know which stack you’d use and links to any similar projects you’ve shipped.