Presentation 2010-10-08
Change Region Detection with Dense Region Matching from Car-Mounted Camera Images
Naoko ENAMI, Norimichi UKITA, Masatsugu KIDODE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose a method to detect scene changes from two images captured the same location at different times and under different capturing conditions by car mounted cameras. In our method, this comparison between two images can be achieved by a two-stage matching algorithm. In the first stage, SIFT keypoints are utilized to obtain the fundamental matrix from the images. Then, in the second stage, a SIFT feature-based dense region matching algorithm (i.e., Graph-cuts) with the epipolar geometry efficiently detects changed regions between the two images. Comparing with a naive SIFT feature-based dense matching and SIFT Flow, our method can result better performance in terms of both computational cost and matching accuracy for images captured by car mounted cameras. The effectiveness is evaluated with a large set of images captured by the car-mounted cameras in real environments.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Car Mounted Camera / Change Region Detection / SIFT / Graph-cuts
Paper # PRMU2010-91,SP2010-47,WIT2010-35
Date of Issue

Conference Information
Committee WIT
Conference Date 2010/10/1(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 Well-being Information Technology(WIT)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Change Region Detection with Dense Region Matching from Car-Mounted Camera Images
Sub Title (in English)
Keyword(1) Car Mounted Camera
Keyword(2) Change Region Detection
Keyword(3) SIFT
Keyword(4) Graph-cuts
1st Author's Name Naoko ENAMI
1st Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology()
2nd Author's Name Norimichi UKITA
2nd Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
3rd Author's Name Masatsugu KIDODE
3rd Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
Date 2010-10-08
Paper # PRMU2010-91,SP2010-47,WIT2010-35
Volume (vol) vol.110
Number (no) 221
Page pp.pp.-
#Pages 6
Date of Issue