An Improved Image Compression Algorithm Using 2D DWT and PCA with Canonical Huffman Encoding.

Area of Science:

• Computer Science
• Digital Image Processing
• Data Compression

Background:

• The increasing volume of digital image data necessitates efficient storage and transmission solutions.
• Existing image compression techniques often face trade-offs between compression ratio, encoding/decoding speed, and image quality.
• Canonical Huffman coding (CHC) offers faster decoding compared to traditional binary Huffman codes, making it attractive for real-time applications.

Purpose of the Study:

• To develop a hybrid image compression model combining Discrete Wavelet Transform (DWT), Canonical Huffman Coding (CHC), and Principal Component Analysis (PCA).
• To enhance image compression efficiency and maintain high visual quality.
• To reduce decoding time compared to existing methods.

Main Methods:

• A lossy compression approach was implemented by applying PCA, followed by DWT and CHC.
• The hybrid model integrates the strengths of DWT, CHC, and PCA for superior performance.
• The proposed technique was evaluated on various color and grayscale images.

Main Results:

• The hybrid method achieved up to 60% compression while preserving high visual quality.
• Significant improvements in Peak Signal-to-Noise Ratio (PSNR) were observed, with color images showing gains of 17 dB and 22 dB, and grayscale images improving by 5 dB and 8 dB.
• The technique reduced bit-per-pixel (bpp) scores, indicating greater compression efficiency.

Conclusions:

• The proposed DWT, CHC, and PCA hybrid model offers a superior image compression solution.
• This method outperforms existing techniques in terms of both PSNR and bpp, providing a better balance between compression and quality.
• The use of CHC contributes to lower decoding times, making the technique suitable for applications requiring fast image reconstruction.