Embeddit: Image Search Using CLIP and LanceDB

Embeddit is a Python/flask app that allows you to search for images using text queries or by uploading an image (image to image search). It uses the OpenAI CLIP for embedding images and stores in the vectors in LanceDB.

Table of Contents

1. Features

2. Prerequisites

3. Setup

   • Setting up a Python Virtual Environment

   • Activating the Virtual Environment

   • Installing Dependencies

4. Usage

5. License

Features

• Text-based image search: Users can enter a text query to search for relevant images.

• Image-based search: Users can upload an image to find visually similar images.

• Efficient similarity search using LanceDB.

• User-friendly web interface for seamless interaction.

• Customizable image folder for indexing and searching.

Prerequisites

Before setting up Embeddit, ensure that you have the following prerequisites installed:

• Python 3.8 or higher

• pip (Python package installer)

Setup

1. navigate to the project folder

   cd Embeddit
   

   create the virtual environment.

   
   python3 -m venv embeddit_env
   
   

   This will create a new virtual environment named embeddit_env.

2. To activate the virtual environment, run the appropriate command based on your operating system:

• For Windows:

  
  embeddit_env\Scripts\activate
  
  

• For macOS and Linux:

  
  source embeddit_env/bin/activate
  
  

3. Install necessary dependencies. I try to keep them at minimum.

   
   pip install -r requirements.txt
   
   

Usage

To run the Embeddit application, follow these steps:

1. Ensure that you have activated the virtual environment.

2. Place the images you want to index and search in the designated image folder (default: images/ folder).

3. Run the following command to start the application:

   
   python app_image_search.py --image-folder path/to/your/image/folder
   
   

   Replace path/to/your/image/folder with the actual path to the folder containing your images. By default, uses images/ folder in the project directory.

4. Open a web browser and visit http://localhost:5000 to access the Embeddit web interface.

5. Use the search bar to enter text queries or upload an image to find visually similar images.

License

Embeddit is released under the MIT License.