Bioinformatics and Medical Applications

4Deep Learning in Gait Abnormality Detection: Principles and Illustrations

Saikat Chakraborty^1,2*, Sruti Sambhavi² and Anup Nandy²

¹School of Computer Engineering, Kalinga Institute of Industrial Technology (KIIT), Bhubaneswar, Odisha, India

²Machine Intelligence Bio-Motion Research Laboratory, Computer Science & Engineering Dept., National Institute of Technology Rourkela, Rourkela, India

Abstract

Cerebral palsy (CP) is a medical condition which is marked by weakened muscle coordination and other dysfunctions. The root cause for this condition is brain damage in prenatal state. This population experiences musculoskeletal disabilities that bring about abnormalities in some form which affect the gait pattern. We propose various machine learning and deep learning techniques including support vector machines (SVM), multilayer perceptron (MLP), Vanilla long short-term memory (LSTM), and Bidirectional LSTM to diagnose CP gait. The gait dataset consists of linear velocity of seven body joints. All the methods have been deployed by taking each time instant as a data point. LSTM demonstrated to be competing in detecting CP gait.

Keywords: LSTM, cerebral palsy 2, deep learning, SVM

4.1 Introduction

Gait analysis is the study of human locomotion, improved via necessary instrumentation for measuring bio mechanics of movements and muscle activities. Gait analysis becomes inevitable in cases where individuals whose ability to walk effectively and safely are affected. The traditional gait ...