Application of Deep learning in Bone age assessment

2018 1st International Conference on Environment Prevention and Pollution Control Technology

生态环境科学

Wang, Yagang^1 ; Zhang, Qianni^1 ; Han, Jungang^1 ; Jia, Yang^1

School of Computer Science, Xi'An University of Posts and Telecommunications, Shaanxi

710121, China^1

关键词: Bone age;    Bone age assessment;    Children and adolescents;    Classification and identifications;    Growth and development;    Hand bones;    Transfer learning;    X-ray image;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/199/3/032012/pdf
DOI  :  10.1088/1755-1315/199/3/032012

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

Bone age assessment is a common method for evaluating the growth and development of children and adolescents. It can be used for diagnosing problems such as rickets and short stature during the growth of adolescents. The traditional bone age assessment method is the doctor's manual treatment of hand bone X-ray images, and comparing with medical standard pictures to achieve bone age assessment. In order to reduce the workload of doctors in identifying X-ray images and the subjective effect of doctors, the method is proposed by combining deep learning with medical images to implement automatic bone age assessment. Different methods are used to segment the hand bone X-ray images and transfer learning for classification and identification of bone age. The test results show that female bone age test precision were assigned 94.4% within 20 months, male test precision were assigned 90.5% within 20 months.