Multimodal Medical Image Fusion in Dual Tree Complex Wavelet Transform Domain Using Maximum and Average Fusion Rules

Getting clinical information from fusion of multimodal medical images is a challenging and emerging area of research. Image fusion is the process of extracting complementary information from different modality of medical image into one single image which will be more beneficial for diagnostics and surgical planning. Use of real valued wavelet transform in fusion suffers from shift sensitivity and poor directionality, therefore, we have used Dual tree complex wavelet transform (DTCWT) for fusion which is approximately shift invariant and has better directionality and proposed a new multilevel multimodal image fusion method using maximum and average fusion rules. To validate our fusion results, we have used two fusion evaluation parameters (mutual information (MI) and Q(AB)(F)) and compared the proposed fusion method with Biorthogonal wavelet transform (BWT), Discrete wavelet transform (DWT) and PCA based image fusion techniques. Quantitative as well as qualitative analysis of the fusion results obtained from the proposed method proved that the proposed method is better than Biorthogonal wavelet transform, Discrete wavelet transform and PCA based image fusion techniques.