Reverse Image Search Engine

This project is a reverse image search engine that leverages the power of a Weaviate vector database to enable users to upload images and find similar images. It is built using Node.js for the backend API and Next.js for the frontend interface.

Features

• Image Upload: Users can upload images to be indexed in the Weaviate vector database.
• Image Search: By uploading an image, users can search for and retrieve similar images from the database.

Technology Stack

• Frontend: Next.js, React, Ant Design UI library
• Backend: Node.js, Express.js
• Database: Weaviate vector database and xata
• Image Processing: Utilizes Weaviate's img2vec-neural module for converting images into vectors and xata for storage and search using elastic search

Prerequisites

• Node.js (LTS version recommended)
• Docker and Docker Compose

How to Run

Start the Weaviate Database

1. Navigate to the home directory and start Weaviate with Docker Compose:

cd Reverse-Image-Search-Engine && docker-compose up -d

Start the Backend API

2. In a new terminal, go to the Embedding-service directory, install dependencies, and start the server:

cd Embedding-service
npm install
npm start

Start the Backend API

3. In a new terminal, go to the Embedding-service directory, install dependencies, and start the server:

cd Vision-service
npm install
npm start

Start the Frontend Application

4. In another terminal, navigate to the client directory, install dependencies, and run the Next.js server:

cd client
npm install
npm run dev

The frontend is now accessible at http://localhost:3000.

Demonstration Video

To see the search engine in action, watch the demonstration video below: