hassaanik/Image_Similarity

Image Similarity Comparison Web Application

This project is a web-based application that allows users to upload images and compare them to find the most visually similar ones based on a user-defined similarity threshold. The application utilizes a pre-trained VGG16 model to extract features from images and measures similarity using cosine similarity.

Features

• Upload an input image and a set of comparison images.
• Define a similarity threshold to control the sensitivity of the comparison.
• Automatically identify and display the most similar images.
• Option to retry with different images or threshold settings.
• User-friendly interface with an impressive, creative design.

Technologies Used

• Backend: Python, Flask, TensorFlow, VGG16, scikit-learn
• Frontend: HTML, CSS, Bootstrap
• Image Processing: PIL, NumPy

Usage

• Upload Images: On the homepage, upload your primary image and a set of comparison images.
• Set Similarity Threshold: Input the similarity threshold value at the start.
• View Results: The application will process the images and display the ones that are most similar to the primary image.
• Try Again: After viewing the results, you have the option to try with different images.

Practical Applications

• E-commerce: Helps in recommending visually similar products to users.
• Digital Asset Management: Assists in organizing and finding similar visual content from a large collection.
• Creative Industry: Useful for artists and designers to find similar visual references.

WebApp