Affiliations: [a] Department of Information Management, National Chung Cheng University, Minhsiung, Chiayi, Taiwan | [b] Advanced Institute of Manufacturing with High-tech Innovations, National Chung Cheng University, Minhsiung, Chiayi, Taiwan E-mail: [email protected]
Abstract: In this study, a novel approach, which is the combination
of wavelet-based edge correlation (WEC), speeded up robust features (SURF)
feature descriptors and feature point matching considering local
constraints, is proposed to register the tissue slices of brains. First, WEC
feature extraction is proposed to effectively extract significant and true
feature points located at the contours and vessels of tissue slices, which
can greatly benefit the subsequent feature point matching. The extracted
feature points are then constructed feature descriptors by means of SURF.
Next, when the matching of feature points is performed, the local
constraints are considered, which can considerably improve the matching
accuracy. Finally, image registration is achieved after the transform is
calculated from the correct image pairs. Compared to original SIFT, original
SURF, improved SURF, and four state-of-the-art approaches, the proposed
method obtains more accurate and faster results for brain image registration
in terms of subjective visual presentation and several registration accuracy
criteria.
Keywords: Wavelet-based edge correlation, speeded up robust features (SURF), local constraints, image registration
