TY - JOUR
T1 - Regression Convolutional Neural Network for Automated Pediatric Bone Age Assessment from Hand Radiograph
AU - Ren, Xuhua
AU - Li, Tingting
AU - Yang, Xiujun
AU - Wang, Shuai
AU - Ahmad, Sahar
AU - Xiang, Lei
AU - Stone, Shaun Richard
AU - Li, Lihong
AU - Zhan, Yiqiang
AU - Shen, DInggang
AU - Wang, Qian
N1 - Funding Information:
Manuscript received July 2, 2018; revised September 11, 2018; accepted October 13, 2018. Date of publication October 19, 2018; date of current version September 4, 2019. This research was supported in part by the Medical Engineering Cross Research Foundation of Shanghai Jiao Tong University under Grant YG2017ZD08, in part by the National Natural Science Foundation of China under Grants 81471733 and 61471390, in part by the National Key R&D Program of China under Grant 2017YFC0107602, and in part by the Science and Technology Commission of Shanghai Municipality under Grant 16410722400. (Corresponding authors: Xiujun Yang, Dinggang Shen and Qian Wang.) X. Ren, L. Xiang, Y. Zhan, and Q. Wang are with the Institute for Medical Imaging Technology, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200030, China (e-mail:,renxuhua@sjtu. edu.cn; xianglei_15@sjtu.edu.cn; yiqiang@gmail.com; wang.qian@sjtu. edu.cn).
Publisher Copyright:
© 2013 IEEE.
PY - 2019/9
Y1 - 2019/9
N2 - Skeletal bone age assessment is a common clinical practice to investigate endocrinology, and genetic and growth disorders of children. However, clinical interpretation and bone age analyses are time-consuming, labor intensive, and often subject to inter-observer variability. This advocates the need of a fully automated method for bone age assessment. We propose a regression convolutional neural network (CNN) to automatically assess the pediatric bone age from hand radiograph. Our network is specifically trained to place more attention to those bone age related regions in the X-ray images. Specifically, we first adopt the attention module to process all images and generate the coarse/fine attention maps as inputs for the regression network. Then, the regression CNN follows the supervision of the dynamic attention loss during training; thus, it can estimate the bone age of the hard (or 'outlier') images more accurately. The experimental results show that our method achieves an average discrepancy of 5.2-5.3 months between clinical and automatic bone age evaluations on two large datasets. In conclusion, we propose a fully automated deep learning solution to process X-ray images of the hand for bone age assessment, with the accuracy comparable to human experts but with much better efficiency.
AB - Skeletal bone age assessment is a common clinical practice to investigate endocrinology, and genetic and growth disorders of children. However, clinical interpretation and bone age analyses are time-consuming, labor intensive, and often subject to inter-observer variability. This advocates the need of a fully automated method for bone age assessment. We propose a regression convolutional neural network (CNN) to automatically assess the pediatric bone age from hand radiograph. Our network is specifically trained to place more attention to those bone age related regions in the X-ray images. Specifically, we first adopt the attention module to process all images and generate the coarse/fine attention maps as inputs for the regression network. Then, the regression CNN follows the supervision of the dynamic attention loss during training; thus, it can estimate the bone age of the hard (or 'outlier') images more accurately. The experimental results show that our method achieves an average discrepancy of 5.2-5.3 months between clinical and automatic bone age evaluations on two large datasets. In conclusion, we propose a fully automated deep learning solution to process X-ray images of the hand for bone age assessment, with the accuracy comparable to human experts but with much better efficiency.
KW - Bone age assessment
KW - deep learning
KW - hand radiograph
KW - regression convolutional neural network
UR - http://www.scopus.com/inward/record.url?scp=85055215069&partnerID=8YFLogxK
U2 - 10.1109/JBHI.2018.2876916
DO - 10.1109/JBHI.2018.2876916
M3 - Article
C2 - 30346295
AN - SCOPUS:85055215069
VL - 23
SP - 2030
EP - 2038
JO - IEEE Journal of Biomedical and Health Informatics
JF - IEEE Journal of Biomedical and Health Informatics
SN - 2168-2194
IS - 5
M1 - 8500181
ER -