An automatic fusion algorithm for multi-modal medical images

Abstract

Multi-modal medical image fusion provides an informative and qualitative image, which enhances the accuracy of clinical diagnosis and surgical planning. The most common and effective approaches for medical image fusion involve principal component analysis (PCA), the discrete wavelet transform and the dual-tree complex wavelet transform (DTCWT). These existing algorithms perform fusion in either the spatial or transform domains. In this paper, an advanced fusion algorithm, which combines the DTCWT with PCA, is proposed to perform fusion on several medical imaging modalities. The input images are decomposed by the DTCWT and different fusion rules are applied to combine the coefficients. While PCA is used to fuse the low frequency coefficients, the high frequency coefficients are fused using the maximum fusion rule. The main advantages of the proposed method are the use of both DTCWT and PCA methods together. The DTCWT extracts the salient information about the input images, and then the PCA method is applied to calculate the principal component based on the information on the input images instead of taking only the average value of the low frequency components. The performance of the proposed method was compared with several existing methods. The experimental results obtained reveal that our proposed fusion algorithm performs better than existing schemes both quantitatively and in terms of visual perception.

Keywords:

• Multi-modal medical image fusion
• PCA
• DWT
• DTCWT
• fusion metrics

Acknowledgements

I would like to thank all the staff of the University of New South Wales, Canberra, Australia and the institution for providing me a research publication fellowship to study at this University. I would also like to acknowledge staff from the Canberra Hospital, Canberra, Australia for providing the experimental data and valuable assistance.

Disclosure statement

No potential conflict of interest was reported by the authors.