원문정보
초록
영어
In this paper, a new spot matching algorithm is proposed. The algorithm compares topological patterns on two central spots to be matched that are selected each from reference and target 2D-PAGE images. Similarity transform is applied to one of two patterns in order to consider and correct global and local distortion before comparison of topological patterns. Matching between neighbor spots from two patterns is performed and similarity is evaluated using the normalized Hausdorff distance(NHD). Finally, matching of spots is determined by the number of neighbor matched pairs, the number of outlier spots and the NHD in turn. The proposed algorithm shows good results, even better matching performance when the relative topology is well preserved and no outlier is included.
목차
1. Introduction
2. Previous Research
3. Proposed Spot Matching Algorithm
3.1. Neighbor Spots
3.2. Algorithm
3.3. Similarity Measure
4. Experimental Results
5. Concluding Remarks
References