Presentation | 2019-09-05 Analysis and Feature Selection of CNN Features Hayato Itoh, Yuichi Mori, Masashi Misawa, Masahiro Oda, Shin-Ei Kudo, Kensaku Mori, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Pathological pattern classification is based on texture patterns in ultra magnified view of polyp surfaces. Deep learning is known as an useful representation learning method with large dataset in several fields including pathological classification of medical images.This representation learning method achieves an optimal representation of patterns for predefined architecture by minimising a value of loss function. However, this is the optimisation in the meaning of maximum likelihood estimation with train data for the given architecture and loss function.Therefore, whether the extracted feature is really discriminative feature or not is unclear. In this work, we analyse discriminative and generalisation ability of deep-learning based feature by comparing with texture future for colorectal endocytoscopic images of polyp surfaces. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Endocytoscopy / automated pathological diagnosis / deep learning / feature selection / manifold learning / definite canonicalisation |
Paper # | PRMU2019-29,MI2019-48 |
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 | ENG-JTITLE |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Analysis and Feature Selection of CNN Features |
Sub Title (in English) | Recognition of Neoplasia by using Endocytoscopic Images |
Keyword(1) | Endocytoscopy |
Keyword(2) | automated pathological diagnosis |
Keyword(3) | deep learning |
Keyword(4) | feature selection |
Keyword(5) | manifold learning |
Keyword(6) | definite canonicalisation |
1st Author's Name | Hayato Itoh |
1st Author's Affiliation | Nagoya University(Nagoya Univ.) |
2nd Author's Name | Yuichi Mori |
2nd Author's Affiliation | Showa University Northern Yokohama Hospital(Showa Univ.) |
3rd Author's Name | Masashi Misawa |
3rd Author's Affiliation | Showa University Northern Yokohama Hospital(Showa Univ.) |
4th Author's Name | Masahiro Oda |
4th Author's Affiliation | Nagoya University(Nagoya Univ.) |
5th Author's Name | Shin-Ei Kudo |
5th Author's Affiliation | Showa University Northern Yokohama Hospital(Showa Univ.) |
6th Author's Name | Kensaku Mori |
6th Author's Affiliation | Nagoya University(Nagoya Univ.) |
Date | 2019-09-05 |
Paper # | PRMU2019-29,MI2019-48 |
Volume (vol) | vol.119 |
Number (no) | PRMU-192,MI-193 |
Page | pp.pp.129-134(PRMU), pp.129-134(MI), |
#Pages | 6 |
Date of Issue | 2019-08-28 (PRMU, MI) |