Prediction of Bone Age in the Wrist Bone Region Based on the TW3-C Carpal Method

Abstract

The Chinese 05 scoring method serves as the established standard for bone age assessment. Bone age assessment is a common task in the field of object detection and classification. In this study, we propose an improved algorithm for bone age detection in the wrist region using the TW3-C Carpal method of the Chinese 05 scoring system. We introduce the application of the improved YOLO V5 algorithm and ViT classification network in bone age prediction in the wrist region and analyze their advantages, including accuracy, recall rate, and mAP (mean average precision). The improved YOLO V5 algorithm and ViT classification network address issues such as poor robustness, inaccurate detection region positioning, and incorrect target category classification that are faced by traditional deep learning algorithms. This improvement enhances the detection accuracy and is of significant importance in accurately predicting bone age.