A single-domain generalization framework integrating Zernike-Based detail blur feature extraction and Mamba global attention mechanism

Key Points

• Enhances generalization performance through a novel framework integrating mechanisms.
• Key metric indicates improved accuracy compared to traditional methods with 15% increase.
• The approach employs Zernike-based detail blur feature extraction alongside global attention techniques.
• Significance highlights the need for effective methods in visual processing tasks across various domains.