Multi-Modal Medical Image Fusion Based On Optimal Mass Transportation Theory

Multi-modal medical image fusion combines medical images from different imaging devices to combine more useful information and delete the redundant information from the source images.In reality,on account of it is impossible to capture all the detail information from one imaging mode.Therefore,the fusion of multi-modal medical images can improve the imaging quality and enhance the clinical applicability of medical images in the diagnosis and evaluation of medical problems for more reliable and accurate evaluation.At present,there are many multi-modal medical image fusion methods,and they have been well applied in clinic treatment.However,they still have many shortcomings,such as too much noise,color distortion of the fusion image and too high complexity of the algorithm,which can not obtain fusion results in real time.In this paper,the research meaning and background of multi-modal medical image fusion are introduced in detail,and the existing fusion algorithms and related theories are summarized and analyzed.For the existing problems,a multi-modal medical image fusion algorithm based on optimal mass transportation theory is proposed for the first time.The mainly research content is as followings:1.A multimodal medical image fusion method based on optimal quality transfer theory is proposed.The image fusion problem is solved from a global perspective.The image fusion problem is expressed as displacement interpolation geodesic line in probability measure space.In order to avoid the influence of black background on fusion results,two source images to be fused are represented as two points in probability measure space by image preprocessing method.Then,the optimal transmission center between two points is calculated by combining the optimal mass transfer theory.In order to improve the calculation efficiency,the original distance is replaced by the loose optimal mass transfer distance.Finally,the inverse process of image preprocessing is used to reconstruct the centroid to obtain the final fusion image.2.At the end of this paper,through detailed experiments,the components of the proposed algorithm are experimented.At the same time,the proposed method is compared with different classical and novelty pixel-level fusion methods from different aspects(subjective,objective and comprehensive score).Finally,the experimental results show that the proposed multi-modal medical image fusion algorithm is effective.The algorithm has good performance and robustness,less running time,and the fused image shows better quality and less noise.