Development of an Automated System for Identification of Skeletal Maturity using Convolutional Neural Networks Approach

Abstract

This study seeks to present an advanced automation skeletal recognition system that receives a radiograph of the left hand, wrist, and fingers as input and outputs a bone age forecast. A Faster R-CNN takes the input of left-hand radiograph produces the detected DRU region from left-hand radiograph. Since the DRU area covers the most of the left-hand area, it helps us to assess the bone maturity of the infants and the juvenile people and predicts the accelerating and retarding phases of puberty. The experimental section contains information on how the 1101 radiographs of the left hand and wrist were set up and how the model performed when various optimization strategies and training sample numbers were applied. After testing parameter adjustment, the suggested system finally achieves 92 percent (radius) and 90 percent (ulna) classification accuracy.