Physical Fatigue Detection From Gait Cycles via a Multi-Task Recurrent Neural Network

Abstract

This paper describes a deep learning approach to classify physically fatigued and nonfatigued gait cycles via a recurrent neural network (RNN), where each gait cycle is represented as a time series of three-dimensional coordinates of body joints. Gait cycles inherently have large intra-class variations caused by gait stance differences (e.g., which foot is supporting/swinging) at the beginning of each gait cycle, which makes it difficult to identify subtle differences induced by fatigue. To overcome these difficulties, we introduce a supporting foot-aware RNN model in a multi-task learning framework for better fatigue detection. More specifically, the RNN model has two branches of layers: one is assigned to the main task of fatigue classification and the other is assigned to the auxiliary task of estimating the first supporting foot in the gait cycles. We collected physically fatigued and non-fatigued gait cycles from eight subjects and conducted experiments to evaluate the accuracies of the proposed multi-task model in comparison to a single-task model. As a result, the proposed method achieved an overall area under curve (AUC) of 0.860 for fatigue classification in a leave-one-subject-out cross-validation, and an AUC of 0.915 in a leave-oneday-out evaluation. It can be concluded from the experimental results that a fatigue detection system for daily use, especially for screening purposes, is very feasible on the basis of the proposed approach.INDEX TERMS Fatigue detection, gait behavior, multi-task learning, recurrent neural network, skeletal joint coordinates

Injury Prevention Effect of MRI Imaging Technology in Physical Education and Sports Training

“…For example, the preparatory activities before training are insufficient and unreasonable, and the coaches ignore the necessity of proper relaxation during the training process and arranged a lot of load sports that do not match the physical fitness level of the trainees, some athletes insisted on training with injuries in order to improve their competition performance and competitive skills, and the coaches did not organize training programs, especially confrontational training programs. Athletes' physical factors limit their athletic ability to a certain extent; if athletes and coaches do not realize this, and arrange training programs that are not in line with their own physiological level, it is very easy to cause sports injuries [ 5 ]. Physiological factors mainly include muscle strength, flexibility, sensitivity, coordination, injury history, and degree of fatigue [ 6 ].…”

Assessment of Cognitive Fatigue from Gait Cycle Analysis

“…GCN considers both the spatial and temporal aspects of gait and achieves state-of-the-art accuracy in the CASIA-B public gait dataset. In a recent study, the authors classified physically fatigued and non-fatigued gait cycles via a multi-task RNN [7]. In the aforementioned study, the proposed model has one primary branch that does the task of fatigue classification while the auxilary branch identifies the first supporting foot in the gait cycles.…”

“…On the other hand, the walking or gait patterns among different subjects have an extremely high variation. This subtle intra-class variation and massive inter-class variation make gait classification a difficult problem for classifiers as the massive inter-class difference overwhelms the subtle inter-class variation [7]. The gait sequence can also be thought of as a temporal sequence of body keypoints.…”

“…Various issues cause disturbance in the gait cycle of an individual such as physical fatigue, cognitive fatigue, heightened emotion, and diseases such as multiple sclerosis, stroke, etc. [7][8][9]. Existing methods of CF detection relies on electroencephalogram (EEG), Functional Magnetic Resonance Imaging (fMRI) data, or eye-motion [10][11][12].…”