Content-Based Image Retrieval System

University of Piraeus | Department of Informatics

BSc course: Image Analysis

Semester: 7

Project Completion Year: 2024

Description

This project implements a Content-Based Image Retrieval (CBIR) system using hypergraph-based algorithm. The system aims to retrieve images from a database based on their visual content rather than metadata. By leveraging advanced techniques such as hypergraph construction and similarity calculations, the project enhances the accuracy and efficiency of image retrieval processes.

Features

• Content-Based Image Retrieval: Retrieval of images based on their content using graph-theoretic algorithms.
• Rank Normalization: Normalization of the ranking list of images for improved accuracy in retrieval.
• Hypergraph Construction: Construction of a hypergraph to represent the relationships between images and their features.
• Hyperedge Similarities Calculation: Calculation of similarities between hyperedges of the hypergraph to evaluate the relationship between images.
• Cartesian Product of Hyperedge Elements: Calculation of the Cartesian product of hyperedge elements to enhance similarity analysis.
• Hypergraph-Based Similarity: Similarity computation based on the constructed hypergraph, allowing for accurate retrieval of relevant images.
• Pre-trained Neural Network Features: Extraction of objective features from pre-trained neural networks (such as SqueezeNet, GoogleNet, ResNet) using intermediate hidden layers.
• Target Image Identification: Ability to designate specific images as target images and present related images from the database.
• Accuracy Measurement: Systematic process for measuring the accuracy of the algorithm and presenting the results.

How to Run

1. Clone the repository:

git clone https://github.com/dimitrisstyl7/image-analysis-project.git

2. Navigate to the project directory:

cd image-analysis-project/Project

3. Create and activate a virtual environment:

On Linux/Mac

python3 -m venv venv
source venv/bin/activate

On Windows

python -m venv venv
venv\Scripts\activate

4. Install dependencies:

pip install -r requirements.txt

5. Run the program:

On Linux/Mac

python3 main.py

On Windows

python main.py

Contributors

Dimitris Stylianou

Panagiota Nicolaou

Niki Dimopoulou

Acknowledgments

This project was developed as part of the "Image Analysis" BSc course at the University of Piraeus. Contributions and feedback are always welcome!

License

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