Tumor images registration based on Chamfer matching algorithm

Clinical diagnosis is often supported by a wide variety of medical images. Different images with their own specialized information complement each other. Biomedical image registration addresses the mapping of two or more images of same patient. In this paper, a Distance Transform (DT) is applied to the edge of the natural brain tissue in the CT image. Then the registration based on Chamfer matching algorithm is accomplished by using the edges of the natural brain tissue in MR-T1 and MR-T2 images as masks. When two images are precisely matched, the fusion of CT and MR is performed. The example used in this paper is the registration of CT and MR images of a brain tumor patient. After matching and fusion, the resultant image shows clear structure for all tissue and distinct contrast between gray matter and white matter. The tumor part in the final image pops out. Also, the details which show how the tumor presses the surrounding brain parenchyma are vivid. The algorithm can also be used as a reference for the matching of other medical images such as SPECT, PET, DSA, etc.