Fashion Recommender System

This project implements a Fashion Recommender System using Convolutional Neural Networks (CNN) and Transfer Learning with ResNET architecture. The system enables reverse image search and groups visually similar products based on extracted image features.

Key Features:

• Feature Extraction: Fine-tuned a ResNET model to extract high-dimensional features from a dataset of 44,000 fashion images.
• Recommendation Generation: Utilized K-Nearest Neighbors (KNN) algorithm to generate recommendations based on Euclidean distance between feature vectors.
• Streamlit Interface: Integrated with a Streamlit interface for user-friendly interaction and visual display of recommendations.

Run Streamlit app:

   streamlit run main.py

Recommendations sample:

The image uploaded by the user is displayed along with the 5 nearest recommendations below it.