AW-ELM-based Crouch Gait recognition after ischemic stroke

Crouch Gait (CG) can be observed in the hemiplegia persons after ischemic stroke. Walking with Crouch Gait (CG) shown a large gaits disorder. This paper explores the use of adaptive wavelet extreme learning machine (AW-ELM) to classifying different gait conditions for hemiplegia and healthy subjects. Three participants having a Crouch Gait problem with categories of Mild, Moderate, and Severe gait conditions, also, one Healthy person are used their data in this work. The recognition system extracting number of time and frequency domain features for dimensionality reduction. While for the classification stage, the common Extreme Learning Machine (ELM) classifiers are used. AW-ELM achieved maximum testing accuracy up to 91.149 % and with using majority vote post-processing the accuracy achieves 91.547 %.