Enhanced Image Compression Pipeline

A Python-based image compression pipeline implemented in a Jupyter Notebook, combining Discrete Wavelet Transform (DWT), K-Means clustering, and Arithmetic Coding for efficient grayscale image compression.

This pipeline allows you to compress images, store them in a lightweight format, and decompress them while preserving visual quality. It also provides metrics like PSNR and SSIM to evaluate compression quality.

---

Features

• Discrete Wavelet Transform (DWT): Reduces image redundancy in frequency space.
• K-Means Clustering: Reduces color intensity variations, lowering data size.
• Arithmetic Coding: Efficient entropy coding for further compression.
• Quality Metrics: Computes PSNR and SSIM for compressed vs decompressed images.
• Visualization: Displays images at each compression stage (original → DWT → clustered → decompressed).
• Pickle-based Storage: Saves compressed data as .pkl files for easy sharing and decompression.

---

Installation

git clone https://github.com/faizmansoor/Image-Compression-Pipeline.git
cd image-compression-pipeline
pip install -r requirements.txt

##Usage

Compress and Decompress an Image

from compression_pipeline import compress_and_decompress_image

Compress and decompress a grayscale image

result = compress_and_decompress_image("test-images/test1.png")

This will:

Save the compressed file as test1_compressed.pkl

Save the decompressed image as test1_decompressed.png

Display images at all stages

Print PSNR, SSIM, and compression ratio

Compression Ratios:

Sample 1: 11.27:1 (91.1% size reduction)

Sample 2: 11.23:1 (91.1% size reduction)

Sample 3: 10.38:1 (90.4% size reduction)

Average Compression Ratio: 10.96:1 (~90.9% size reduction)

Images