Verify Kicks

Verify Kicks is a prototype sneaker authentication platform. A buyer uploads photos of a pair, a computer vision model scores authenticity signals across key regions, and the result comes back as a confidence-scored verification rather than a binary verdict.

• ·Sneaker resale buyers pay large premiums with no fast, affordable way to check authenticity before purchase.
• ·Expert authentication is slow, manual and priced per item, so most buyers skip it entirely.
• ·Marketplaces carry the counterfeit risk with no scalable verification layer of their own.

• ·Framed the product around the moment of purchase decision: verification has to fit inside minutes, not days.
• ·Built an image-upload flow where users photograph defined sneaker regions for consistent model input.
• ·Trained a YOLO-based detection model to locate and score authenticity signals on uploaded images.
• ·Served inference through a FastAPI backend, with a labelled real-vs-counterfeit dataset pipeline for continuous model improvement.
• ·Designed a dual business model: B2C per-check verification plus a B2B verification API for resale marketplaces.

• ✓Working prototype: upload to scored verification result end to end
• ✓Dataset pipeline in place so every submission improves the model
• ✓Two validated distribution paths - per-check consumer checks and a marketplace API
• ✓Prototype stage - not yet publicly launched, no performance claims made