Elite Wear — Virtual Try-On Fashion Store

Static fashion catalogs offer limited ways to explore how a garment might look on a shopper. This DEPI training project explored an interactive clothing-preview concept, rather than a production fitting or checkout service.

The team built a fashion-store prototype. I developed the homepage, About and Contact pages, project documentation and the experimental try-on app. It uses webcam input, TensorFlow.js and PoseNet to estimate upper-body landmarks, then draws an adjustable T-shirt image using Canvas. Other storefront components were delivered by the team.

HTML, CSS and JavaScript power the interface. TensorFlow.js loads PoseNet in the browser; shoulder and hip positions inform the Canvas overlay dimensions. The try-on code does not use a custom backend AI service. Source inspection verifies the implementation; live webcam playback has not been verified in this review environment. Manual overlay controls and keypoint-based placement remain experimental, with alignment and realism limitations.

My role: homepage, About page, Contact page, README and the Try-On application. The repository credits Mohamed for these contributions; the wider storefront is team work.

Implemented flow: open the try-on page → permit camera access → load the pose-estimation model → estimate upper-body keypoints → draw the T-shirt overlay → adjust the preview → capture a PNG. This describes the source implementation, not a completed live-camera verification.

Limitations: this is a pose-guided 2D garment overlay, not photorealistic clothing generation, accurate sizing or fit prediction. Account and checkout screens are prototypes; a production authentication or payment backend has not been established.

Publication review: attribution, image rights and an authentic webcam demonstration remain to be approved. The reported completion year is 2024; the exact completion date is not confirmed.