Orientation adaptive subband coding of images.

Area of Science:

• Digital image processing
• Signal processing
• Information theory

Background:

• Conventional subband coding schemes exhibit Gibbs-like phenomena at image edges, especially at low bit rates.
• The directionality of local image features remains largely unexploited in existing subband coding techniques.

Purpose of the Study:

• To develop a novel subband coding scheme that leverages the orientation of local image features.
• To mitigate Gibbs-like artifacts in reconstructed images.
• To improve subjective image quality compared to conventional methods.

Main Methods:

• An orientation adaptive subband coding scheme was proposed.
• The scheme utilizes information about the orientation of local image features.
• Performance was evaluated against conventional separable subband coding and JPEG compression.

Main Results:

• The orientation adaptive scheme effectively avoids objectionable Gibbs-like phenomena at reconstructed image edges.
• At comparable bit rates, subjective image quality is considerably enhanced.
• The proposed scheme shows superior performance over conventional separable subband coding and JPEG.

Conclusions:

• Exploiting local image feature orientation in subband coding significantly improves image reconstruction quality.
• The orientation adaptive scheme offers a viable solution for high-quality image compression at low bit rates.