| Presentation | 2019-09-05 Hierarchical Classification to Detect Type of Diseases and Abnormality Simultaneously in Optical Coherence Tomography Images Yudai Kato, Yuji Ayatsuka, Takaki Uta, Soichiro Kuwayama, Hideaki Usui, Aki Kato, Yuichiro Ogura, Tsutomu Yasukawa, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Analyzing medical images with machine learning is useful not only for classifying types of diseases but for screening abnormality. Our previous work showed that a convolutional neural network (CNN) model which learned for classifying diseases detects abnormality better than a CNN model which just learned abnormality as one category. The result is regarded as that a type of disease is important information to find visual feature of abnormality in image. In this paper, we propose a hierarchical method in which a model is trained both types of diseases and abnormality simultaneously. In our method, losses for each diseases are used for training the lower layer, and a loss for abnormality calculated as simple accumulation of losses for each diseases is used for training the upper layer. Models trained by our method achieve better accuracy in both classifying diseases and screening. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | OCT / fundus diseases / machine learning |
| Paper # | PRMU2019-26,MI2019-45 |
| Date of Issue | 2019-08-28 (PRMU, MI) |
| Conference Information | |
| Committee | PRMU / MI / IPSJ-CVIM |
|---|---|
| Conference Date | 2019/9/4(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Yoichi Sato(Univ. of Tokyo) / Yoshiki Kawata(Tokushima Univ.) |
| Vice Chair | Toru Tamaki(Hiroshima Univ.) / Akisato Kimura(NTT) / Takayuki Kitasaka(Aichi Inst. of Tech.) / Hidekata Hontani(Nagoya Inst. of Tech.) |
| Secretary | Toru Tamaki(NTT) / Akisato Kimura(OMRON SINICX) / Takayuki Kitasaka(Yamaguchi Univ.) / Hidekata Hontani(Univ. of Hyogo) |
| Assistant | Yusuke Uchida(DeNA) / Takayoshi Yamashita(Chubu Univ.) / Hotaka Takizawa(Tsukuba Univ.) / Yoshito Otake(NAIST) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding / Technical Committee on Medical Imaging / 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) | Hierarchical Classification to Detect Type of Diseases and Abnormality Simultaneously in Optical Coherence Tomography Images |
| Sub Title (in English) | |
| Keyword(1) | OCT |
| Keyword(2) | fundus diseases |
| Keyword(3) | machine learning |
| Keyword(4) | |
| Keyword(5) | |
| 1st Author's Name | Yudai Kato |
| 1st Author's Affiliation | CRESCO LTD.(CRESCO) |
| 2nd Author's Name | Yuji Ayatsuka |
| 2nd Author's Affiliation | CRESCO LTD.(CRESCO) |
| 3rd Author's Name | Takaki Uta |
| 3rd Author's Affiliation | CRESCO LTD.(CRESCO) |
| 4th Author's Name | Soichiro Kuwayama |
| 4th Author's Affiliation | Nagoya City University(Nagoya City University) |
| 5th Author's Name | Hideaki Usui |
| 5th Author's Affiliation | Nagoya City University(Nagoya City University) |
| 6th Author's Name | Aki Kato |
| 6th Author's Affiliation | Nagoya City University(Nagoya City University) |
| 7th Author's Name | Yuichiro Ogura |
| 7th Author's Affiliation | Nagoya City University(Nagoya City University) |
| 8th Author's Name | Tsutomu Yasukawa |
| 8th Author's Affiliation | Nagoya City University(Nagoya City University) |
| Date | 2019-09-05 |
| Paper # | PRMU2019-26,MI2019-45 |
| Volume (vol) | vol.119 |
| Number (no) | PRMU-192,MI-193 |
| Page | pp.pp.105-108(PRMU), pp.105-108(MI), |
| #Pages | 4 |
| Date of Issue | 2019-08-28 (PRMU, MI) |