6 Medical Image Fusion Schemes using Contourlet Transform and PCA Based

Fusion imaging is one of the most modern, accurate and useful diagnostic techniques in medical imaging today. The new technology has made a clear difference in patient care by compressing the time between diagnosis and treatment. Although image fusion can have different purposes, the main aim of fusion is spatial resolution enhancement or image sharpening. Also known as integrated imaging, it provides a computer link that allows for the combination of multimodal medical images into a single image with more complete and accurate description of the same object. The benefits are even more profound in combining anatomical imaging modalities with functional ones. For example, PET-CT in lung cancer, MRI-PET in brain tumors, SPECT-CT in abdominal studies and ultrasound images-MRI for vascular blood flow (Patias, 2002). Outcome of MRI-CT image fusion has been shown to be able to assist in planning surgical procedure. Mainly, medical image fusion try to solve the issue of where there is no single modality provides both anatomical and functional information. Further more information provided by different modalities may be in agreement or in complementary nature. An important research issue in medical image processing, specifically in information computation, is fusion of multimodal information (Daneshvar and Ghassemian, 2007; Hong et al., 2007; Zhongfei et al., 2003). Existing algorithms generally use discrete wavelet transform (DWT) (Anna et al., 2007; Pajares and Manuel de la Cruz, 2004; Singh et al., 2009) for multimodal medical image fusion because DWT preserves different frequency information in stable form and allows good localization both in time and spatial frequency domain. However, one of the major drawbacks of DWT is that the transformation does not provide shift invariance. This causes a major change in the wavelet coefficients of the image even for minor shifts in the input image. In medical imaging, it is important to know and preserve the exact location of this information, but shift variance may lead to inaccuracies. As an alternative (Kingsbury, 1999), proposed dual-tree complex wavelet transform (DT-CWT) which provides approximate shift invariance. DT-CWT has the drawback of limited directional information. Hence, contourlet transform was proposed to capture the most important salient information in images by incorporating the DT-CWT and DFB (Chen and Li, 2005).