AI Clothing Virtual Try-On Development

I’m adding an end-to-end virtual fitting room to my e-commerce site so visitors can upload a full-body photo and immediately see our apparel draped on their own figure—accurate pose, lighting, and cloth dynamics included. The workflow I have in mind combines human-parsing, garment warping, and a lightweight generative model that streams results straight to the browser.

Core features I need implemented:
• Real-time try-on with latency low enough to feel instant
• Save & share so customers can download a composite or post directly to social platforms
• Size recommendations that read body proportions from the uploaded image and match them to our inventory’s size chart

I’ll provide product photography and size tables; you’ll build the computer-vision back-end (segmentation, pose estimation, garment transfer, fit prediction) and the front-end widget that plugs into a React storefront. Python + PyTorch for the model and a WebGL/WebRTC layer for rendering are preferred, but I’m open to alternatives if they hit the same performance marks.

Deliverables are:
1. A containerised API that receives a full-body JPEG/PNG and a garment SKU and returns the composited image plus size suggestion.
2. Embeddable JavaScript/TypeScript component with hooks for upload, preview, save, and share.
3. Technical documentation and a short video demo showing real-time performance on desktop and mobile.
4. Deployment assistance to push the service into my AWS stack.

Acceptance is based on visual realism (no major artifacts at 1080p), sub-2-second response for typical images, and correct size prediction on our test set.

If you have shipped similar CV or generative-AI experiences, especially for clothing, I’d love to see a link or repo. Show More