Interdisciplinary Approaches to Image Processing for Medical Robotics

C.h. Leung Stephen^*

• Department of Engineering, The University of HongKong, HongKong, China

The advancement of medical robotic systems highlights the critical need for precise and high-quality visual data, particularly in low-quality imaging scenarios. This study explores the interdisciplinary physics underlying image fusion and analysis, addressing challenges such as integrating complementary features, handling dynamic range variations, and suppressing noise in real-world medical contexts. We introduce the Multi-Scale Feature Adaptive Fusion Network (MFAFN) and the Dynamic Feature Refinement Strategy (DFRS), which leverage principles from computational and experimental physics to enhance imaging techniques. MFAFN applies multiscale feature extraction, attention-based alignment, and adaptive fusion to improve spatial and spectral integration while preserving crucial details. Complementing this, DFRS employs saliencybased weighting, context-aware mechanisms, and dynamic normalization to refine feature importance and mitigate inconsistencies. This interdisciplinary approach bridges computational physics, non-linear systems, and technological development, delivering significant improvements in fusion quality metrics such as spatial consistency, edge retention, and noise suppression. Our findings contribute to advancing medical robotics by integrating novel physical principles into imaging methodologies, supporting sustainable innovations in healthcare technology.