Joint Chroma Subsampling and Distortion-Minimization-Based Luma Modification for RGB Color Images With Application

Area of Science:

• Digital Image Processing
• Color Image Reconstruction
• Video Compression

Background:

• Chroma subsampling is a common technique in image and video compression.
• Existing methods may introduce quality distortions in reconstructed full-color images.
• Improving the fidelity of reconstructed images, especially in video coding, remains a challenge.

Purpose of the Study:

• To propose a novel hybrid method for improving the quality of reconstructed RGB full-color images.
• To minimize color peak signal-to-noise ratio (CPSNR) distortion in reconstructed images.
• To adapt the method for digital time delay integration and Bayer mosaic images.

Main Methods:

• A hybrid approach combining conventional chroma subsampling (4:2:0) with distortion-minimization-based luma modification.
• Development of a main theorem to determine ideally modified luma blocks in constant time.
• Optimization of the method for specific image types like digital time delay integration and Bayer mosaic images.

Main Results:

• Substantial quality improvement in reconstructed RGB images was demonstrated.
• Enhanced performance in terms of CPSNR, visual effect, and CPSNR-bitrate trade-off.
• The method showed superior Bjøntegaard delta PSNR performance compared to existing schemes.

Conclusions:

• The proposed hybrid method effectively improves reconstructed RGB image quality.
• The method offers a globally optimal solution for minimizing color distortion.
• The technique is adaptable and beneficial for various image types and video coding applications.