Multiscale medical image fusion in wavelet domain

Area of Science:

• Medical Imaging
• Signal Processing
• Computer Vision

Background:

• Medical image fusion is crucial for enhancing diagnostic accuracy.
• Existing fusion techniques, particularly in the wavelet domain, face challenges in achieving optimal results.
• Multimodal medical image analysis requires robust fusion strategies.

Purpose of the Study:

• To propose a novel multiscale fusion method for multimodal medical images in the wavelet domain.
• To enhance the flexibility and relevance of fused medical images through multi-scale analysis.
• To rigorously evaluate the proposed method against existing state-of-the-art fusion techniques.

Main Methods:

• A multiscale fusion approach in the wavelet domain was developed.
• Maximum selection rule was applied for fusion at multiple scales (minimum to maximum level).
• The method was experimentally validated using diverse medical image datasets.

Main Results:

• The proposed method demonstrated superior fusion performance compared to pyramid and wavelet-based methods, including Principal Component Analysis (PCA).
• Subjective and objective evaluations confirmed the effectiveness of the multiscale fusion approach.
• Key metrics such as edge strength (Q), mutual information (MI), entropy (E), standard deviation (SD), blind structural similarity index metric (BSSIM), spatial frequency (SF), and average gradient (AG) were used for comparison.

Conclusions:

• The proposed multiscale wavelet domain fusion method is effective for multimodal medical images.
• The approach offers enhanced flexibility and improved fusion quality, aiding in medical image analysis.
• The findings support the utility of this method for improving diagnostic capabilities in medical imaging.