A Wearable Smartphone-Enabled Camera-Based System for Gait Assessment
Abstract
Quantitative assessment of gait parameters, i.e., step length (SL), step width (SW), step time (ST), gait speed (GS) and their variability, provide valuable diagnostic and prognostic information. This is more useful if data is collected during normal daily activities. Most gait analysis systems are bulky, expensive, need special operation and designed to be used in laboratory settings. Recently, wearable systems have attracted considerable attention due to their lower cost, portability which enables to assess not only laboratory setting but also real world data. Here, a simple wearable smartphone-enabled system for measurement of spatiotemporal gait parameters (SmartGait) is introduced. This system uses a smartphone camera (attached to the waist with belt holster), two circular fiducial markers of known diameter (4.65cm), and a 90-degree wide-angle lens, to capture gait parameters. SmartGait measures SL, SW, ST, gait speed and their variability using the data extracted from the video images and on-board inertial sensors of the smartphone (Apple, iPhone 5s, iOS7). The video images are processed to isolate and locate the foot markers, which are used to calibrate each video frame (sampling rate 60 Hz). Trunk motion as measured by the embedded inertial sensor (sample rate 100 Hz) is used for image stabilization. Predefined angular distortion coefficient matrix is also used to correct angular distortion. SmartGait was compared for concurrent agreement against two laboratory-based systems, i.e., a motion analysis system (OptoTrak) and an instrumented walkway (GaitRite, using fifteen healthy young adults (mean 25.8, S.D. 2.6 years) walking at slow, preferred, and fast speed. For the comparing different data set from three systems, intra-class correlation coefficients (ICC) show good to excellent concurrent agreement between SmartGait and the other systems (ICCs between 0.73 and 0.93 for on-board assessment and 0.83 to 0.97 for corrected off-board assessment).
Degree
M.S.E.C.E.
Advisors
Ziaie, Purdue University.
Subject Area
Computer Engineering|Electrical engineering|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.