Enhancing gait pattern analysis with deep learning on image data

Abstract

Current techniques for analysing gait patterns from picture data lack the sturdiness and precision necessary for real-world clinical applications and extensive research. It is necessary to create and test deep learning models that can efficiently improve gait pattern analysis by taking advantage of differences in subject appearances and ambient variables while extracting useful characteristics from photos. This study looks at and develops deep learning methods for accurate and affordable image-based gait pattern analysis in an effort to solve this issue. In this paper we have proposed deep convolutional neural network (DCNN) that gives best accuracy of 97.80 % and error of 0.04 % thanLong Short-Term Memory Model (LSTM), Radial Basis Functional Neural Network and LinearNeural Network (LNN) on image data. This model will be helpful in clinical healthcare to detect human gait pattern with better accuracy and less error.