Browse Prior Art Database

ANALYSIS OF GAIT PATTERNS TO PREDICT NEUROLOGICAL DISORDERS

IP.com Disclosure Number: IPCOM000240573D
Publication Date: 2015-Feb-10
Document File: 7 page(s) / 534K

Publishing Venue

The IP.com Prior Art Database

Related People

S. Margret Anouncia: INVENTOR [+3]

Abstract

A training-testing paradigm has been proposed in this paper to envisage the task of classifying an abnormal gait pattern into one of the five types viz. Parkinsonian, Scissor, Spastic, Steppage and Normal gait. It takes into account the motion and posture features for two cases viz. right-leg-front and left-leg-front. Unlike usage of sensors and other devices, this work exploits the visual cues of a gait pattern to assess the nature of abnormality for a gait pattern. A Gaussian Mixture Model has been trained for the classification task. The application of such a system lies in prediction of certain neurological diseases that are attributed to abnormal gait patterns.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 18% of the total text.

Page 01 of 7

7

Abstract: A training-testing paradigm has been proposed in this paper to envisage the task of classifying an abnormal gait pattern into one of the five types viz. Parkinsonian, Scissor, Spastic, Steppage and Normal gait. It takes into account the motion and posture features for two cases viz. right-leg-front and left-leg-front. Unlike usage of sensors and other devices, this work exploits the visual cues of a gait pattern to assess the nature of abnormality for a gait pattern. A Gaussian Mixture Model has been trained for the classification task. The application of such a system lies in prediction of certain neurological diseases that are attributed to abnormal gait patterns.

---------- ----------

INTRODUCTION

Gait has been defined as coordinated cyclic combination of movements resulting in human locomotion. Examples of gait are walking, jogging, climbing steps etc. Gait possesses a potential to recognize individuals at a distance and at a low resolution, mostly in surveillance areas where other biometrics like fingerprint recognition and iris recognition can't play the role. According to the article entitled "MedlinePlus" by NIH the causes of specific gait patterns e.g. Propulsive gait (causes: carbon monoxide poisoning, Parkinson's disease), Spastic gait (brain or head trauma, brain tumor, cerebral palsy), Steppage gait (causes: multiple sclerosis, spinal cord injury) etc. According to literature there exist three approaches for abnormal gait analysis are image processing, floor sensors and sensors placed on body. This piece of work focuses on exploiting visual cues associated with abnormalities of certain gait patterns under training and testing paradigm.

SYSTEM OVERVIEW

The proposed method encompasses the following phases for accomplishment of abnormal gait analysis and classification under a training-testing paradigm.

1

ANALYSIS OF GAIT PATTERNS TO PREDICT NEUROLOGICAL DISORDERS.

Kuhelee Roy, G Subrahmanya VRK Rao , S. Margret Anouncia.

Fig 1. System overview of analysis of gait pattern



Page 02 of 7

7

2

CONTRIBUTION:

The method proposed in this paper for analyzing and classifying abnormal is basically three-fold:


1. Tracking of the upper and lower part of body across video frames.

2. Analyzing the shape and motion patterns between two phases' viz. right-leg-front (RLF) and left-leg- front (LLF) (segmented by means of active contour)

3. Assessing the degree of abnormality in terms of the extracted features against trained features for each of the gait patterns considered in this paper viz. Parkinsonian gait, Scissor gait, Spastic gait, Steppage gait.

RESULTS


i.Classification of abnormal gait pattern

Given a video exhibiting an abnormal gait, the motion and posture pattern of the subject in the video is tested against the trained motion and posture features. As shown in figure, the posture pattern of most of the frames of a given video (Normal gait) is closer to that of trained object for Normal gait posture, thereby conclu...