NSCT-based multimodal medical image fusion using pulse-coupled neural network and modified spatial frequency

Area of Science:

• Medical Imaging
• Image Processing
• Artificial Intelligence

Background:

• Multimodal medical image fusion (MIF) is crucial for enhancing diagnostic accuracy by combining complementary information from different imaging modalities.
• Existing MIF techniques often face challenges in preserving fine details and reducing artifacts.
• The integration of advanced signal processing and neural network techniques offers potential for improved fusion performance.

Purpose of the Study:

• To present a novel multimodal medical image fusion (MIF) method.
• To leverage the strengths of non-subsampled contourlet transform (NSCT) and pulse-coupled neural network (PCNN) for superior fusion outcomes.
• To demonstrate the effectiveness of the proposed method in generating high-quality fused medical images.

Main Methods:

• The proposed method utilizes non-subsampled contourlet transform (NSCT) for image decomposition.
• Low-frequency subbands are fused using a 'max selection' rule.
• High-frequency subbands are fused using a pulse-coupled neural network (PCNN) motivated by modified spatial frequency.
• Inverse NSCT (INSCT) is applied to reconstruct the fused image.

Main Results:

• The proposed NSCT-PCNN based MIF method demonstrated effective fusion of multimodal medical images.
• Subjective and objective analyses confirmed the superiority of the proposed method compared to state-of-the-art techniques.
• The fusion process successfully preserved important image features and reduced noise.

Conclusions:

• The developed multimodal medical image fusion technique based on NSCT and PCNN is effective.
• This method offers significant advantages for medical image analysis and diagnosis.
• The proposed approach represents a valuable contribution to the field of medical image processing.