AUTHOR=Zhao Yangyang , Zhou Jiali , Qiu Fei , Liao Xuying , Jiang Jianhua , Chen Heqing , Lin Xiaomei , Hu Yiqun , He Jianquan , Chen Jian 

TITLE=A deep learning method for foot-type classification using plantar pressure images

JOURNAL=Frontiers in Bioengineering and Biotechnology

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/bioengineering-and-biotechnology/articles/10.3389/fbioe.2023.1239246

DOI=10.3389/fbioe.2023.1239246

ISSN=2296-4185

ABSTRACT=Foot pressure distribution can be used as a quantitative parameter for evaluating the anatomical deformity of the foot and diagnosing and treating pathological gait and falling. In this paper, we proposed a deep learning (DL) model for predicting foot type based on the pressure distribution of the foot. We analyzed 1572 plantar images by using four networks, namely YOLO v5, improved YOLO-v5 with SE, improved YOLO-v5 with CBAM attention, and a multilabel classification model based on ResNet-50, to obtain a suitable model for foot type recognition. Plantar pressure images of 46 subjects were used to validate and compare these models. The multilabel classification algorithm based on ResNet-50 outperformed the other algorithms. The improved YOLO-v5 model with SE, the improved YOLO-v5 model with CBAM attention, and the multilabel classification model based on ResNet-50 achieved accuracies of 0.652, 0.717, and 0.826, respectively, significantly higher than those obtained using the ordinary plantar-pressure system and the standard YOLO-v5 model. These results indicate that the proposed DL-based multilabel classification model based on ResNet-50 is superior in flat foot type detection and can be used to evaluate the clinical rehabilitation status of patients with abnormal foot types and various foot pathologies when more data on patients with various diseases are available for training.