DCMamba: A differential coupling network based on dynamic convolution and mamba for misaligned image registration and fusion

Key Points

• Image registration accuracy improves with DCMamba, highlighting its effectiveness in aligning misaligned images.
• With a 20% improvement over traditional methods, DCMamba demonstrates significant advancements in image fusion techniques.
• Assessment using differential coupling and dynamic convolution methods results in superior alignment for various imaging modalities.
• This framework may enable enhanced applications in medical imaging, though further validation is necessary for clinical use.