Presentation 2018-09-21
Multi-modal image registration based on local mutual information
Hiroaki Tani, Toshikazu Wada,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In recent years, various types of images having different modality from visible-light images, such as infrared images, depth images, and so on, are widely used for specific tasks. Compared with visible light images, these images of the same scene have different intensities and gradient directions. Such features are not suitable for image registration. A representative multi-modal image registration proposed so far is based on mutual information between pixel values. This method maximizes mutual information between pixel values of two images. However, this method fails for image pairs having completely different modalities, such as thermal and visible-light images. This is because the pixel values between two images do not have one-to-one correspondence. For solving this problem, we propose an image registration method that maximizes ‘local mutual information’, which is the sum of mutual information values computed for local image blocks. This method is based on the observation that the one-to-one pixel value correspondence can stand for local image blocks. Through image registration experiments between visible-light image and FIR/depth images, we confirmed that our method works better than the mutual-information based method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Local mutual information / multi-modal images / registration / mutual information
Paper # PRMU2018-65,IBISML2018-42
Date of Issue 2018-09-13 (PRMU, IBISML)

Conference Information
Committee PRMU / IBISML / IPSJ-CVIM
Conference Date 2018/9/20(2days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Shinichi Sato(NII) / Hisashi Kashima(Kyoto Univ.)
Vice Chair Yoshihisa Ijiri(Omron) / Toru Tamaki(Hiroshima Univ.) / Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo)
Secretary Yoshihisa Ijiri(NEC) / Toru Tamaki(Osaka Univ.) / Masashi Sugiyama(Nagoya Inst. of Tech.) / Koji Tsuda(AIST)
Assistant Go Irie(NTT) / Yoshitaka Ushiku(Univ. of Tokyo) / Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.)

Paper Information
Registration To Technical Committee on Pattern Recognition and Media Understanding / Technical Committee on Infomation-Based Induction Sciences and Machine Learning / Special Interest Group on Computer Vision and Image Media
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Multi-modal image registration based on local mutual information
Sub Title (in English)
Keyword(1) Local mutual information
Keyword(2) multi-modal images
Keyword(3) registration
Keyword(4) mutual information
1st Author's Name Hiroaki Tani
1st Author's Affiliation Wakayama University(Wakayama Univ)
2nd Author's Name Toshikazu Wada
2nd Author's Affiliation Wakayama University(Wakayama Univ)
Date 2018-09-21
Paper # PRMU2018-65,IBISML2018-42
Volume (vol) vol.118
Number (no) PRMU-219,IBISML-220
Page pp.pp.207-212(PRMU), pp.207-212(IBISML),
#Pages 6
Date of Issue 2018-09-13 (PRMU, IBISML)