Multimodal medical image fusion using adaptive co-occurrence filter-based decomposition optimization model.

Area of Science:

• Medical image processing
• Computational imaging

Background:

• Medical image fusion merges information from multiple images for enhanced diagnosis and surgical planning.
• Existing fusion methods often lack optimal information complementarity and efficiency.

Purpose of the Study:

• To propose a novel adaptive co-occurrence filter (ACOF)-based image decomposition optimization model for superior medical image fusion.
• To enhance the quality and information content of fused medical images.

Main Methods:

• Introduced the concept of skewness of pixel intensity (SPI) and applied it to design the ACOF.
• Developed an optimization model for image decomposition, replacing traditional iterative filters for better base and detail layer separation.
• Generated fused images using designed fusion rules and linear addition.

Main Results:

• The proposed ACOF method significantly improved fused image quality compared to 22 state-of-the-art methods.
• Achieved superior performance across five objective indices and subjective evaluations.
• Demonstrated higher computational efficiency than existing fusion techniques.

Conclusions:

• The novel SPI and ACOF-based optimization model offers a significant advancement in medical image fusion.
• The method provides high-quality fused images beneficial for clinical diagnosis and surgical planning.
• The approach is computationally efficient and outperforms current standards.