Improved JPEG Coding by Filtering 8 × 8 DCT Blocks.

Area of Science:

• Digital image processing
• Data compression algorithms
• Information theory

Background:

• JPEG is a widely adopted standard for image compression, utilizing lossy and lossless techniques.
• Current JPEG entropy coding (Huffman, arithmetic) processes all 8x8 DCT blocks, including empty ones, using end-of-block markers.
• These markers for empty blocks increase file size unnecessarily, impacting storage and transmission efficiency.

Purpose of the Study:

• To propose a modified JPEG entropy coding method to enhance compression ratios.
• To address the inefficiency of handling empty blocks in the standard JPEG compression process.
• To reduce file size without compromising image quality.

Main Methods:

• Developed a modified JPEG entropy coding approach.
• Implemented a system to store empty block locations in a separate buffer.
• Compressed the location buffer using an efficient lossless method.
• Evaluated coding performance using bits per pixel and peak signal-to-noise ratio (PSNR).

Main Results:

• The modified algorithm achieves a higher compression ratio compared to standard JPEG methods.
• The proposed method results in lower bits per pixel for test images.
• Image quality, measured by PSNR, is maintained at acceptable levels.

Conclusions:

• The modified JPEG entropy coding effectively reduces file size by optimizing the handling of empty blocks.
• This approach offers a more efficient image compression solution within the JPEG framework.
• The method provides a favorable trade-off between file size reduction and image fidelity.