摘要

Medical image registration is a very important technology in the field of medical image fusion, in order to meet the needs of clinical diagnosis and treatment. This paper presents a registration method of medical images based on contour features and multi-scale information. First, the resample operator is used to reduce the resolution of the couple images, and it can reduce the amount of computation effectively in the rough registration process. Second, Canny operator and method of scan line are used to extract the feature contour of the reduced dimension image. Then the principal axis method is applied to realize the rough registration based on contour lines. Finally, the subband images obtained by wavelet transform are used to accomplish the hierarchical registration. The experimental results show that the proposed method can guarantees the accuracy of registration and also reduces the amount of computation in the process of registration, and as well as in the condition of the image has a partial incomplete, we can also get the ideal registration results.