imgemb 0.2.5

A lightweight Python library for generating image embeddings with semantic search

imgemb: Efficient Image Embeddings and Similarity Search

imgemb is a powerful Python library for generating image embeddings and performing similarity search. It offers multiple embedding methods, semantic search capabilities, and interactive visualizations.

Features

• 🎨 Multiple Embedding Methods:

  • Average Color: Fast color-based similarity
  • Grid: Spatial color distribution analysis
  • Edge: Shape and structure detection
  • Semantic: CLIP-based content understanding

• 🔍 Similarity Search:

  • Fast nearest neighbor search
  • Multiple distance metrics (cosine, euclidean)
  • Batch processing support
  • Interactive result visualization

• 🖼️ Interactive Visualization:

  • Plot similar images with scores
  • Save interactive HTML plots
  • Customizable layouts and titles

• 🛠️ Command Line Interface:

  • Generate embeddings
  • Compare images
  • Find similar images
  • Semantic search

• 🚀 Performance:

  • Efficient numpy-based computations
  • GPU support for semantic search
  • Optimized for large image collections

Installation

pip install imgemb

Quick Start

from imgemb import ImageEmbedder, plot_similar_images

# Initialize embedder
embedder = ImageEmbedder(method="grid")

# Find similar images
similar_images = embedder.find_similar_images(
    "query.jpg",
    "images/",
    top_k=5
)

# Visualize results
fig = plot_similar_images("query.jpg", similar_images)
fig.show()

Command Line Usage

# Generate embeddings
imgemb generate images/ --output embeddings.json --method grid

# Find similar images
imgemb find-similar query.jpg images/ -k 5 --method edge

# Semantic search
imgemb search "a photo of a dog" images/ -k 5

Documentation

• Usage Guide: Detailed examples and best practices
• API Reference: Complete API documentation
• Examples: Sample scripts and notebooks

Contributing

We welcome contributions! Please see our Contributing Guide for details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use imgemb in your research, please cite:

@software{imgemb2024,
  author = {Aryan Raj},
  title = {imgemb: Efficient Image Embeddings and Similarity Search},
  year = {2024},
  publisher = {GitHub},
  url = {https://github.com/aryanraj2713/imgemb}
}