Multimodality Medical Image Registration Using Harris Corner And Mutual Information

Medical imaging has been an important part of modern medical sciences. Medical imaging includes two separate parts: medical imaging system and medical image processing. Imaging system lays stress on the image acquiring process. Processing is focused on image post-processing in order to improve image quality, extract image features or fuse different model of images. With the development of imaging technology, the modality medical images have been used widely in clinical diagnoses and surgical therapies. Integrating those images is helpful to improve the accuracy of clinical diagnoses and surgical therapies. Image registration is the key part of integration; it is a geometrical transformation and can rectify two according to the corresponding features. Image registration research is the emphases of this dissertation.This thesis firstly introduces the conception major registration algorithms and categories are and process of registration. According to whether extracting images' feature or not, we divide present methods into two parts: feature extraction based methods and vowel similarity based methods. We then described the knowledge about the affection of interpolation. Compared the three interpolation algorithms of nearest neighbor, linear and PV (Partial Volume), experiment results indicated that PV interpolation gave the best performance in terms of the smoothness of mutual information function and the precision of registration. Mutual information registration method thinks that the mutual information of the wrapped part of the images reaches the maximum when two images register best. Compared to conventional registration methods, the predominance of Mutual information registration method is that it doesn't need to suppose the relation of the images and that it need no image segmentation and image preprocess, and that it can bring to success automatically without alternation with people. Due to the inferiority of image registration using maximum mutual information, we researched the registration based on corner points. We used Harris operator in consideration of repetition and information measure. Mutual information was proposed to be the similarity measure of corner points.Medical images registrations usually use local optimizing algorithms, and the global optimizing algorithms are less concerned. But these local algorithms are easy to fall into the local supreme value, so it will lead to the wrong registration. The paper uses genetic algorithm which has good global searching ability. Because of the earliness, slow searching speed and long running time, genetic algorithm has some defects when it is applied in medical images registration. The paper improves the standard genetic algorithm from coding, genetic operating, and so on. It improves the capability of the algorithm.Given the head images characteristic of MRI and pet, we present a new registration strategy: we adopt MI (Mutual Information of Harris corner points) as cost function, PV as interpolation, GA as our optimization algorithm to search the best registration parameters. Theories indicated that this method have the advantages of high precision , good robust.