|
|||||
|
|
Automated Classification of Epiphyses in the Distal Radius and Ulna using a Support Vector Machine |
|
| Authors: | Ya‐hui Wang MD Tai‐ang Liu PhD Hua Wei MD Lei Wan MD Chong‐liang Ying BA Guang‐you Zhu MD |
| Affiliation: | 1. Institute of Forensic Science, Ministry of Justice, China, Shanghai Key Laboratory of Forensic Medicine, NO1347 West Guangfu Road, Putuo District, Shanghai, China;2. School of Materials Science and Engineering, Shanghai University, NO99 Shangda Road, Baoshan District, Shanghai, China;3. Institute of Shanghai Huayi Forensic Science, NO 1277 Dingxi Road, Changning District, Shanghai, China;4. Additional information and reprint requests:;5. Guang‐you ZHU, M.D;6. Shanghai Key Laboratory of Forensic Medicine;7. Institute of Forensic Science, Ministry of Justice;8. 200063 Shanghai;9. China;10. E‐mail: |
| Abstract: | The aim of this study was to automatically classify epiphyses in the distal radius and ulna using a support vector machine (SVM) and to examine the accuracy of the epiphyseal growth grades generated by the support vector machine. X‐ray images of distal radii and ulnae were collected from 140 Chinese teenagers aged between 11.0 and 19.0 years. Epiphyseal growth of the two elements was classified into five grades. Features of each element were extracted using a histogram of oriented gradient (HOG), and models were established using support vector classification (SVC). The prediction results and the validity of the models were evaluated with a cross‐validation test and independent test for accuracy (PA). Our findings suggest that this new technique for epiphyseal classification was successful and that an automated technique using an SVM is reliable and feasible, with a relative high accuracy for the models. |
| Keywords: | forensic science forensic anthropology distal radius distal ulna support vector classification support vector machine histogram of oriented gradient |
