TY - GEN
T1 - TOCSAC
T2 - 5th International Symposium on Advances in Visual Computing, ISVC 2009
AU - He, Zhoucan
AU - Wang, Qing
AU - Yang, Heng
PY - 2009
Y1 - 2009
N2 - This paper aims at outliers screening for the feature correspondence in image matching. A novel robust matching method, called topology constraint sample consensus (TOCSAC), is proposed to speed up the matching process while keeping the matching accuracy. The TOCSAC method comprises of two parts, the first of which is the constraint of points order, which should be invariant to scale, rotation and view point change. The second one is a constraint of affine invariant vector, which is also used to validate in similar and affine transforms. Comparing to the classical algorithms, such as RANSAC (random sample consensus) and PROSAC (progressive sample consensus), the proposed TOCSAC can significantly reduce time cost and improve the performance for wide base-line image correspondence.
AB - This paper aims at outliers screening for the feature correspondence in image matching. A novel robust matching method, called topology constraint sample consensus (TOCSAC), is proposed to speed up the matching process while keeping the matching accuracy. The TOCSAC method comprises of two parts, the first of which is the constraint of points order, which should be invariant to scale, rotation and view point change. The second one is a constraint of affine invariant vector, which is also used to validate in similar and affine transforms. Comparing to the classical algorithms, such as RANSAC (random sample consensus) and PROSAC (progressive sample consensus), the proposed TOCSAC can significantly reduce time cost and improve the performance for wide base-line image correspondence.
UR - https://www.scopus.com/pages/publications/72449156195
U2 - 10.1007/978-3-642-10520-3_58
DO - 10.1007/978-3-642-10520-3_58
M3 - 会议稿件
AN - SCOPUS:72449156195
SN - 364210519X
SN - 9783642105197
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 608
EP - 618
BT - Advances in Visual Computing - 5th International Symposium, ISVC 2009, Proceedings
Y2 - 30 November 2009 through 2 December 2009
ER -