A Novel Algorithm Based on SIFT and Graph Transformation for Mammogram Registration
Mammogram registration is an important step in the processing of automatic detection of breast cancer. It provides aid to better visualization correspondence on temporal pairs of mammograms. This paper presents a novel algorithm based on SIFT feature and
- PDF / 184,286 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 99 Downloads / 188 Views
A Novel Algorithm Based on SIFT and Graph Transformation for Mammogram Registration Yang-jun Zhong and Lan-Zhen Chen
Abstract Mammogram registration is an important step in the processing of automatic detection of breast cancer. It provides aid to better visualization correspondence on temporal pairs of mammograms. This paper presents a novel algorithm based on SIFT feature and Graph Transformation methods for mammogram registration. First, features are extracted from the mammogram images by scale invariant feature transform (SIFT) method. Second, we use graph transformation matching (GTM) approach to obtain more accurate image information. At last, we registered a pair of mammograms using Thin-Plate spline (TPS) interpolation based on corresponding points on the two breasts, and acquire the mammogram registration image. Performance of the proposed algorithm is evaluated by three criterions. The experimental results show that our method is accurate and closely to the source images. Keywords Thin-plate spline • SIFT • Graph transformation • Mammogram registration
243.1
Introduction
Image registration is the process of overlaying two or more images of the same scene taken at different times, from different viewpoints, and/or by different sensors [1]. Research in the area of image registration has been receiving considerable attention, and a number of algorithms have been proposed over the last
Y.-j. Zhong (*) Jiangxi University of Science and Technology, Ganzhou, China e-mail: [email protected] L.-Z. Chen Gannan Medical University, Ganzhou, China e-mail: [email protected] S. Zhong (ed.), Proceedings of the 2012 International Conference on Cybernetics 1897 and Informatics, Lecture Notes in Electrical Engineering 163, DOI 10.1007/978-1-4614-3872-4_243, # Springer Science+Business Media New York 2014
1898
Y.-j. Zhong and L.-Z. Chen
10 years. These algorithms of registration are applied to remote sensing images [2], natural scenes images [3], and some forensic science images [4], etc. Rueckert et al. employed a free-form deformation model based on B-splines to describe the local deformation of 3D breast MR images [5]. Jianzhe et al. proposed the free-form deformation based on non-uniform rational B-splines (NURBS) to acquire non-rigid transformation [6]. In contrast, the feature-based methods are more accurate and faster to compute as long as the algorithms of feature extraction are reliable. Urschler et al. presented a feature-based non-linear registration method consist of 3D corner detection, local SIFT feature descriptor and global shape context feature descriptor, robust feature matching and calculation of a dense displacement field [7]. The primary sources of difference are variations in positioning, compression and changes normally encountered in breast. So, precise mammogram registration is much difficult and intractable [8].
243.2
Extracting Control Points Pairs
Medical images present in the mammogram database often contain deliberately inserted identifiable labels, which need to be eliminated
Data Loading...
