This research introduces a novel approach to medical image translation that preserves anatomical structures with significantly higher accuracy than previous methods.

• Addresses the critical problem of structure distortion in synthesized medical images
• Implements dynamic frequency balance to prevent overfitting high-frequency information
• Uses knowledge guidance to maintain anatomical integrity during translation
• Enables more precise cross-modality image synthesis for enhanced clinical diagnosis

Why it matters: Accurate anatomical structures in synthesized medical images are essential for reliable diagnosis. This approach could improve diagnostic accuracy and reduce the need for multiple imaging procedures, saving time and resources in clinical settings.

Structure-Accurate Medical Image Translation based on Dynamic Frequency Balance and Knowledge Guidance