Presentation 2023-09-08
[Short Paper] Construction of Cell Nucleus Classifier using Complementary-Label Learning towards the Quantification of Grading for Follicular Lymphoma
Ryoichi Koga, Mauricio Kugler, Tatsuya Yokota, Kouichi Ohshima, Hiroaki Miyoshi, Miharu Nagaishi, Noriaki Hashimoto, Ichiro Takeuchi, Hidekata Hontani,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we report the cell type classification from a pathological image toward the subtype classification of malignant lymphoma. Follicular lymphoma, one of the subtypes, is divided into several grades based on the number of a certain cell, called a centroblast. However, in clinical practice, the grade is qualitatively evaluated without recognizing each type of a large number of nuclei. Although a cell nucleus classifier is required to construct the quantitative criteria of the grade, giving each nucleus a label indicating its type is quite difficult even for expert pathologists. Hence, we propose a method to construct the classifier using complementary-label learning with effectively utilizing labels that contain ambiguity. In this paper, we demonstrate the effectiveness of our proposed method and the considerations toward the quantification of the grade.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Image classification / Ambiguous labels / Complementary-label learning / Quantitative evaluation criteria / Malignant lymphoma
Paper # MI2023-14
Date of Issue 2023-09-01 (MI)

Conference Information
Committee MI
Conference Date 2023/9/8(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Ryo Haraguchi(Univ. of Hyogo)
Vice Chair Hideaki Haneishi(Chiba Univ.) / Takayuki Kitasaka(Aichi Inst. of Tech.)
Secretary Hideaki Haneishi(Yamaguchi Univ.) / Takayuki Kitasaka(NAIST)
Assistant Takeshi Hara(Gifu Univ.) / Kenichi Morooka(Okayama Univ.)

Paper Information
Registration To Technical Committee on Medical Imaging
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) [Short Paper] Construction of Cell Nucleus Classifier using Complementary-Label Learning towards the Quantification of Grading for Follicular Lymphoma
Sub Title (in English)
Keyword(1) Image classification
Keyword(2) Ambiguous labels
Keyword(3) Complementary-label learning
Keyword(4) Quantitative evaluation criteria
Keyword(5) Malignant lymphoma
1st Author's Name Ryoichi Koga
1st Author's Affiliation Nagoya Institute of Technology(NIT)
2nd Author's Name Mauricio Kugler
2nd Author's Affiliation Nagoya Institute of Technology(NIT)
3rd Author's Name Tatsuya Yokota
3rd Author's Affiliation Nagoya Institute of Technology(NIT)
4th Author's Name Kouichi Ohshima
4th Author's Affiliation Kurume University(Kurume Univ.)
5th Author's Name Hiroaki Miyoshi
5th Author's Affiliation Kurume University(Kurume Univ.)
6th Author's Name Miharu Nagaishi
6th Author's Affiliation Kurume University(Kurume Univ.)
7th Author's Name Noriaki Hashimoto
7th Author's Affiliation RIKEN(RIKEN)
8th Author's Name Ichiro Takeuchi
8th Author's Affiliation Nagoya University(Nagoya Univ.)
9th Author's Name Hidekata Hontani
9th Author's Affiliation Nagoya Institute of Technology(NIT)
Date 2023-09-08
Paper # MI2023-14
Volume (vol) vol.123
Number (no) MI-185
Page pp.pp.1-2(MI),
#Pages 2
Date of Issue 2023-09-01 (MI)