| Presentation | 2019-03-04 Variational Bayes algorithm of region base coupled MRF with hidden phase variables Naoki Wada, Masaichiro Mizumaki, Yoshiki Seno, Masato Okada, Akai Ichiro, Toru Aonishi, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | There are two methods in coupled Markov Random Field(MRF) model for image segmentation: edge-based method and region-based method. Region-based method is easier to implement and more robust to the noise than edge-based method. However, region-based method is often trapped to local optima. We focus on region base coupled MRF with hidden phase variables, which is reported to be less likely to be trapped to local optima. This model has already been implemented in a LSI circuit but has not been constructed probabilistic inference algorithm and evaluated its performance. We derive fast approximate inference algorithm using variational Bayes and compare with the conventional Ising spin model. Then, we have found that the hidden phase variables model is less likely to be trapped at local optima than the conventinal method. Moreover, we compare the variational method with Markov chain Monte Carlo to verify approximation accuracy of the variational method. Furthermore, we conducted mesoscopic structure detection as an application of image segmentation. Hidden phase variables model provides us with good result in real data. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | MRF / Variational Bayes / MCMC / image segmentation |
| Paper # | NC2018-59 |
| Date of Issue | 2019-02-25 (NC) |
| Conference Information | |
| Committee | NC / MBE |
|---|---|
| Conference Date | 2019/3/4(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | University of Electro Communications |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Yutaka Hirata(Chubu Univ.) / Masaki Kyoso(TCU) |
| Vice Chair | Hayaru Shouno(UEC) / Taishin Nomura(Osaka Univ.) |
| Secretary | Hayaru Shouno(Nagoya Univ.) / Taishin Nomura(NAIST) |
| Assistant | Keiichiro Inagaki(Chubu Univ.) / Takashi Shinozaki(NICT) / Takumi Kobayashi(YNU) / Ｙａｓｕｙｕｋｉ Ｓｕｚｕｋｉ(Osaka Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Neurocomputing / Technical Committee on ME and Bio Cybernetics |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Variational Bayes algorithm of region base coupled MRF with hidden phase variables |
| Sub Title (in English) | |
| Keyword(1) | MRF |
| Keyword(2) | Variational Bayes |
| Keyword(3) | MCMC |
| Keyword(4) | image segmentation |
| 1st Author's Name | Naoki Wada |
| 1st Author's Affiliation | Tokyo Institute of Technology(Tokyo Inst. of Tech.) |
| 2nd Author's Name | Masaichiro Mizumaki |
| 2nd Author's Affiliation | Japan Synchrotron Radiation Research Institute(JASRI) |
| 3rd Author's Name | Yoshiki Seno |
| 3rd Author's Affiliation | Saga prefectural regional industry support center(Saga prefectural regional industry support center) |
| 4th Author's Name | Masato Okada |
| 4th Author's Affiliation | The University of Tokyo(The Univ. of Tokyo) |
| 5th Author's Name | Akai Ichiro |
| 5th Author's Affiliation | Kumamoto University(Kumamoto Univ.) |
| 6th Author's Name | Toru Aonishi |
| 6th Author's Affiliation | Tokyo Institute of Technology(Tokyo Inst. of Tech.) |
| Date | 2019-03-04 |
| Paper # | NC2018-59 |
| Volume (vol) | vol.118 |
| Number (no) | NC-470 |
| Page | pp.pp.87-92(NC), |
| #Pages | 6 |
| Date of Issue | 2019-02-25 (NC) |