Back to projects
Fashion Recommendation System
PrototypeA visual similarity search engine for fashion products: it turns catalog images into ResNet50 embeddings and returns the most visually similar items with nearest-neighbor search, wrapped in a Streamlit app.
- Python
- ResNet50
- Streamlit
- Computer Vision
- Machine Learning
A visual similarity search engine for fashion products. Instead of matching on tags or keywords, it looks at the image itself: each product photo is passed through a ResNet50 network to produce an embedding, and a nearest-neighbor search returns the items that look most alike. The whole thing is served through a Streamlit app so you can upload or pick an item and see similar products back.