Presentation | 2013-12-12 Robust Image Matching based on Probabilistic Estimation of Disturbed Pixels using Analysis of Intensity Variation Masataka SAITO, Manabu HASHIMOTO, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this paper, we propose an image matching method which is robust for sudden disturbance such as occlusion. As for the disturbance by occlusion, it is difficult to predict it in advance, which is a different point to disturbance by illumination changes. In this research, pixels affected by disturbance are estimated in matching process, and influence of unexpected disturbance will be reduced by removing these pixels from similarity calculation. By using statistical variation analysis, if there are pixels that have unnatural intensity compared with statistical changes of illumination intensity, they are estimated as disturbed pixels. This is how a matching which is robust for occlusion problem will be realised. In addition, we have achieved high-speed computation of estimating disturbed pixels by reducing the number of reference pixels. Recognition success rate of 82.2% and processing time of 413msec was achieved for experiments using 1048 actual images including partially occluded target object. We proved that recognition rate of the proposed method is approximately 10% higher than a conventional method, the Selective correlation coefficient. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | image matching / pixel selection / occlusion / robustness / analysis of disturbed pixels |
Paper # | PRMU2013-75 |
Date of Issue |
Conference Information | |
Committee | PRMU |
---|---|
Conference Date | 2013/12/5(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Pattern Recognition and Media Understanding (PRMU) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Robust Image Matching based on Probabilistic Estimation of Disturbed Pixels using Analysis of Intensity Variation |
Sub Title (in English) | |
Keyword(1) | image matching |
Keyword(2) | pixel selection |
Keyword(3) | occlusion |
Keyword(4) | robustness |
Keyword(5) | analysis of disturbed pixels |
1st Author's Name | Masataka SAITO |
1st Author's Affiliation | Graduate School of Information Science and Technology, Chukyo University() |
2nd Author's Name | Manabu HASHIMOTO |
2nd Author's Affiliation | Graduate School of Information Science and Technology, Chukyo University |
Date | 2013-12-12 |
Paper # | PRMU2013-75 |
Volume (vol) | vol.113 |
Number (no) | 346 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |