Multi-modal Medical Image Fusion Based on Geometric Algebra Discrete Cosine Transform

Multi-modal medical image fusion refers to the combination of patient area images obtained under diverse or identical imaging modalities, which improves the clinical applicability and provides more specific disease information for diagnosis. However, most of the existing image fusion algorithms usually divided color images into three channels of R, G, B for processing separately, which ignores the correlation between the channels and easily causes image information loss and blurring. This paper proposes a multi-modal color medical image fusion algorithm based on geometric algebra discrete cosine transform (GA-DCT). The GA-DCT algorithm combines the character of GA, which represents the multi-vector signal as a whole, can improve the quality of the fusion image and avoid a large number of complex operations related to encoding and decoding. Firstly, the source images are divided into several image blocks and expressed in GA multi-vector form; Secondly, we extend the traditional DCT to GA space and propose GA-DCT; Thirdly, we use GA-DCT to decompose the image to obtain AC and DC coefficients and finally a fusion algorithm are used to fuse the images. The experimental results show that the proposed algorithm can get clear and comprehensive fusion image, which also has great advantages under different compression ratios.