Deep Bayesian active learning-to-rank with relative annotation for estimation of ulcerative colitis severity

被引:3
作者
Kadota, Takeaki [1 ]
Hayashi, Hideaki [2 ]
Bise, Ryoma [1 ,3 ]
Tanaka, Kiyohito [4 ]
Uchida, Seiichi
机构
[1] Kyushu Univ, Dept Adv Informat Technol, 744,Motooka,Nishi Ku, Fukuoka, Fukuoka 8190395, Japan
[2] Osaka Univ, Inst Databil Sci, 2-8,Yamadaoka, Suita, Osaka 5650871, Japan
[3] Natl Inst Informat, Res Ctr Med Bigdata, 2-1-2,Hitotsubashi,Chiyoda Ku, Tokyo 1018430, Japan
[4] Kyoto Second Red Cross Hosp, Dept Gastroenterol, 355-5,Haruobicho Kamigyo Ku, Kyoto, Kyoto 6028026, Japan
基金
日本科学技术振兴机构;
关键词
Computer-aided diagnosis; Learning to rank; Active learning; Relative annotation; Endoscopic image dataset; DISEASE; INFORMATION; IMAGES;
D O I
10.1016/j.media.2024.103262
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Automatic image-based severity estimation is an important task in computer-aided diagnosis. Severity estimation by deep learning requires a large amount of training data to achieve a high performance. In general, severity estimation uses training data annotated with discrete (i.e., quantized) severity labels. Annotating discrete labels is often difficult in images with ambiguous severity, and the annotation cost is high. In contrast, relative annotation, in which the severity between a pair of images is compared, can avoid quantizing severity and thus makes it easier. We can estimate relative disease severity using a learning-to-rank framework with relative annotations, but relative annotation has the problem of the enormous number of pairs that can be annotated. Therefore, the selection of appropriate pairs is essential for relative annotation. In this paper, we propose a deep Bayesian active learning-to-rank that automatically selects appropriate pairs for relative annotation. Our method preferentially annotates unlabeled pairs with high learning efficiency from the model uncertainty of the samples. We prove the theoretical basis for adapting Bayesian neural networks to pairwise learning-to-rank and demonstrate the efficiency of our method through experiments on endoscopic images of ulcerative colitis on both private and public datasets. We also show that our method achieves a high performance under conditions of significant class imbalance because it automatically selects samples from the minority classes.
引用
收藏
页数:12
相关论文
共 43 条
[1]   The power of ensembles for active learning in image classification [J].
Beluch, William H. ;
Genewein, Tim ;
Nuernberger, Andreas ;
Koehler, Jan M. .
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, :9368-9377
[2]  
Burges Chris., 2005, ICML 05 P 22 INT C M, P89
[3]  
Carterette B., 2006, Proceedings of the Twenty-Ninth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, P629, DOI 10.1145/1148170.1148289
[4]   Automated classification of gastric neoplasms in endoscopic images using a convolutional neural network [J].
Cho, Bum-Joo ;
Bang, Chang Seok ;
Park, Se Woo ;
Yang, Young Joo ;
Seo, Seung In ;
Lim, Hyun ;
Shin, Woon Geon ;
Hong, Ji Taek ;
Yoo, Yong Tak ;
Hong, Seok Hwan ;
Choi, Jae Ho ;
Lee, Jae Jun ;
Baik, Gwang Ho .
ENDOSCOPY, 2019, 51 (12) :1121-1129
[5]  
Dasgupta S., 2008, P 25 INT C MACHINE L, P208
[6]  
Gal Y, 2016, Arxiv, DOI arXiv:1506.02157
[7]  
Gal Y, 2017, PR MACH LEARN RES, V70
[8]  
Gal Y, 2016, PR MACH LEARN RES, V48
[9]  
Gorriz Marc, 2017, 31 C MACHINE LEARNI
[10]   Training and deploying a deep learning model for endoscopic severity grading in ulcerative colitis using multicenter clinical trial data [J].
Gutierrez Becker, Benjamin ;
Arcadu, Filippo ;
Thalhammer, Andreas ;
Gamez Serna, Citlalli ;
Feehan, Owen ;
Drawnel, Faye ;
Oh, Young S. ;
Prunotto, Marco .
THERAPEUTIC ADVANCES IN GASTROINTESTINAL ENDOSCOPY, 2021, 14