AI Somatotype System Using 3D Body Images: Based on Deep-Learning and Transfer Learning DOI Creative Commons
Jiwun Yoon, Sang‐Yong Lee, Ji-Yong Lee

и другие.

Applied Sciences, Год журнала: 2024, Номер 14(6), С. 2608 - 2608

Опубликована: Март 20, 2024

Humans share a similar body structure, but each individual possesses unique characteristics, which we define as one’s type. Various classification methods have been devised to understand and assess these types. Recent research has applied artificial intelligence technology utilizing noninvasive measurement tools, such 3D scanner, minimize physical contact. The purpose of this study was develop an somatotype system capable predicting the three types proposed by Heath-Carter’s theory using images collected scanner. To classify types, measurements were taken determine components (endomorphy, mesomorphy, ectomorphy). MobileNetV2 utilized transfer learning model. results are follows: first, AI model showed good performance, with training accuracy around 91% validation 72%. respective loss values 0.26 for set 0.69 set. Second, model’s performance test data resulted in accurate predictions 18 out 21 new points, prediction errors occurring cases, indicating approximately 85% accuracy. This provides foundational subsequent aiming predict 13 detailed across Furthermore, it is hoped that outcomes can be practical settings, enabling anyone smartphone camera identify various based on captured obesity diseases.

Язык: Английский

Evaluating waist-to-hip ratio in youth using frequency-modulated continuous wave radar and machine learning DOI Creative Commons

Jun Byung Park,

Jin Joo Choi, Jae Yoon Na

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 31, 2025

Язык: Английский

Процитировано

0
AI Somatotype System Using 3D Body Images: Based on Deep-Learning and Transfer Learning DOI Creative Commons
Jiwun Yoon, Sang‐Yong Lee, Ji-Yong Lee

и другие.

Applied Sciences, Год журнала: 2024, Номер 14(6), С. 2608 - 2608

Опубликована: Март 20, 2024

Humans share a similar body structure, but each individual possesses unique characteristics, which we define as one’s type. Various classification methods have been devised to understand and assess these types. Recent research has applied artificial intelligence technology utilizing noninvasive measurement tools, such 3D scanner, minimize physical contact. The purpose of this study was develop an somatotype system capable predicting the three types proposed by Heath-Carter’s theory using images collected scanner. To classify types, measurements were taken determine components (endomorphy, mesomorphy, ectomorphy). MobileNetV2 utilized transfer learning model. results are follows: first, AI model showed good performance, with training accuracy around 91% validation 72%. respective loss values 0.26 for set 0.69 set. Second, model’s performance test data resulted in accurate predictions 18 out 21 new points, prediction errors occurring cases, indicating approximately 85% accuracy. This provides foundational subsequent aiming predict 13 detailed across Furthermore, it is hoped that outcomes can be practical settings, enabling anyone smartphone camera identify various based on captured obesity diseases.

Язык: Английский

Процитировано

1