Digital Affine Shear Filter Banks with 2-Layer Structure and Their Applications in Image Processing.

Area of Science:

• Digital Signal Processing
• Image Processing
• Applied Mathematics

Background:

• Filter banks are crucial for signal and image processing.
• Affine shear transforms offer unique properties for data analysis.
• Existing frame-based transforms have limitations in efficiency and performance.

Purpose of the Study:

• To construct and analyze digital affine shear filter banks with a 2-layer structure.
• To establish the perfect reconstruction property of these filter banks.
• To demonstrate their advantages in image processing applications.

Main Methods:

• Development of digital affine shear filter banks (DAS-2) with a 2-layer structure.
• Implementation using transition and subdivision operators.
• Redundancy rate analysis and numerical experiments for image processing.

Main Results:

• DAS-2 filter banks exhibit the perfect reconstruction property.
• Redundancy rate is no more than 8 and decreases with more directional filters.
• Demonstrated superior performance compared to state-of-the-art frame-based transforms in image processing.

Conclusions:

• The study successfully constructed and validated DAS-2 filter banks.
• Established a connection between DAS-2 filter banks and 2-layer affine shear tight frames.
• The proposed filter banks offer an efficient and effective tool for advanced image processing.