Shelf Match is a counter assistant for a small skincare shop with a short shelf. The shopper takes a webcam photo or uploads a JPEG or PNG. Shelf Match sends that photo through Perfect Corp's YouCam AI Skin Analysis API and reads the returned scores. A higher ui_score means YouCam rated that area healthier. Shelf Match then applies shop-readable rules: a concern is flagged when ui_score is below 75, a bottle is recommended only when it is in stock and tagged for a flagged concern, and a bottle is listed under leave-on-the-shelf when tagged avoid for a flagged concern. Each card shows the arithmetic. If a flagged concern has no in-stock tag, the shopper is told the shelf cannot help with it.
The catalog is sixteen made-up example products (labeled example data). The store-owner page lets the owner edit helps/avoid tags and in-stock, and highlights concerns with no in-stock product. Privacy: the selfie is processed in memory and not written to disk; in live mode bytes go to YouCam then are dropped on this server.
Live demo: https://shelf-match-production.up.railway.app Source: https://gitlab.com/cryptobandit/shelf-match
Built With
- fastapi
- python
- railway
- youcam-ai-skin-analysis
Log in or sign up for Devpost to join the conversation.