All issues
Current issue
Displaying 1-1 of 1 articles from this issue
- |<
- <
- 1
- >
- >|
Paper
-
Shinya SEKIKAWA, Kazuki IWATA, Keisuke KON2024 Volume 48 Issue 1 Pages 36-43
Published: 2024
Released on J-STAGE: March 20, 2024
JOURNAL FREE ACCESSThis study aimed to develop a measurement system that calculates the body dimension values required for wheelchair fitting by only taking photographs. We first obtained twenty-six key point coordinates using a pre-trained model to detect five anatomical landmarks. Then, we predicted the landmark coordinates using deep learning (neural network: NN) and linear regression (LR) with the keypoint coordinates. Finally, we evaluated the detection accuracy using the coefficient of determination and the mean absolute error between the predicted and true values. The results showed that NN models were more accurate than LR models in both indices and that each landmark coordinate could be detected with an error of less than 20 mm in any NN model. This study demonstrated the feasibility of using NN for wheelchair fitting measurement and suggested its potential clinical application.View full abstractDownload PDF (2546K)
- |<
- <
- 1
- >
- >|