###A Comparative Analysis of Dimensionality Reduction Techniques
##Project Overview This project is an interactive web-based platform designed to explore and apply three advanced image compression techniques: Principal Component Analysis (PCA), Singular Value Decomposition (SVD), and Independent Component Analysis (ICA).
Built using Python and Flask, the application allows users to upload images, select a compression method, and adjust the number of components via an interactive slider to observe real-time trade-offs between image quality and file size.
##Research & Methodology This application is based on a formal study titled "Optimizing Image Compression: A Comparative Analysis of Dimensionality Reduction Techniques".
The study evaluated performance across different image resolutions (256x256, 512x512, and 1024x1024) using three key metrics:
- MSE (Mean Squared Error): Quantifies the difference between pixels.
- PSNR (Peak Signal-to-Noise Ratio): Measures reconstruction quality.
- Compression Ratio: Evaluates storage efficiency.
##Key Research Findings
| Technique | Best For | Conclusion |
|---|---|---|
| PCA | High Fidelity | Consistently offers the best image reconstruction quality with the highest PSNR. |
| SVD | Efficiency | Provides superior compression ratios, ideal for prioritizing storage and bandwidth. |
| ICA | Feature Analysis | Advantageous for applications requiring the extraction of independent components. |
##Features
- Interactive UI: Real-time adjustment of components using a slider.
- Side-by-Side Comparison: View the original and compressed images simultaneously.
- Live Metrics: Real-time calculation of PSNR values to quantify quality loss.
- Multi-Algorithm Support: Toggle between PCA, SVD, and ICA instantly.
##Tech Stack
- Backend: Python (Flask)
- Mathematical Libraries: NumPy, Scikit-learn, OpenCV
- Frontend: HTML5, CSS3, JavaScript
- Documentation: Full research paper included in /docs
##Project StructurePlain
├── app.py # Main Flask application & Logic
├── static/ # CSS and Image assets
├── templates/ # HTML Frontend
├── docs/ # Research paper and documentation (PDF)
├── requirements.txt # List of dependencies
└── README.md # Project overview
##Installation & Usage
- Clone the repository: git clone https://github.com/khairsahil/Image-Compression-Analysis.git
- Install dependencies: pip install -r requirements.txt
- Run the app: python app.py
- Open your browser and navigate to http://127.0.0.1:5000.
Developed as part of the M.Sc. Data Science curriculum at Kirti M. Doongursee College.