A Robust Medical Image Registration Algorithm Based on the SAM of Multi-Scale Harris Corners

To make up for the lack of concern on spatial information in conventional mutual information based image registration framework, this paper designs a novel registration algorithm based on the SAM information of multi-scale Harris corners (CSAM for short). First, the multi-scale contour is extracted, and multi-scale Harris corner detector is added to acquire the estimated transform parameters; and then CSAM is regarded as Similarity Measure function, several optimized match points are obtained, the finally registration parameters are resolved by using least squares method. This algorithm realizes registration of medical images with noise and multiresolutions, further more, it only matches corners and doesn’t need optimal searching, so it has reduced calculate time and avoided local extremum. Experimental results on clinical CT and T1-weighted MR images demonstrate that, as compared with the conventional mutual information based method, the proposed method consistently completes much higher precision, faster speed and better robustness. Keywords-image registration; square root arithmetic mean divergence(SAM); Harris corner; multi-scale

[1]  Max A. Viergever,et al.  Image registration by maximization of combined mutual information and gradient information , 2000, IEEE Transactions on Medical Imaging.

[2]  A. Hero,et al.  Image matching using alpha-entropy measures and entropic graphs , 2005 .

[3]  Charles V. Stewart,et al.  Keypoint Descriptors for Matching Across Multiple Image Modalities and Non-linear Intensity Variations , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[4]  Colin Studholme,et al.  An overlap invariant entropy measure of 3D medical image alignment , 1999, Pattern Recognit..

[5]  Max A. Viergever,et al.  Normalized mutual information based registration using k-means clustering and shading correction , 2006, Medical Image Anal..

[6]  Daniel Rueckert,et al.  Non-rigid registration using higher-order mutual information , 2000, Medical Imaging.

[7]  Takeo Kanade,et al.  The Role of Non-Overlap in Image Registration , 2005, IPMI.

[8]  A. Murat Tekalp,et al.  Local Image Registration by Adaptive Filtering , 2006, IEEE Transactions on Image Processing.

[9]  Mateu Sbert,et al.  High-Dimensional Normalized Mutual Information for Image Registration Using Random Lines , 2006, WBIR.

[10]  Jürgen Weese,et al.  A comparison of similarity measures for use in 2-D-3-D medical image registration , 1998, IEEE Transactions on Medical Imaging.

[11]  Max A. Viergever,et al.  Mutual-information-based registration of medical images: a survey , 2003, IEEE Transactions on Medical Imaging.

[12]  Julien Ros,et al.  Description of Local Singularities for Image Registration , 2006, 18th International Conference on Pattern Recognition (ICPR'06).