Automatic gait event detection in pathological gait using an auto-selection approach

FONSECA ; DUMAS ; ARMAND

Type de document
COMMUNICATION AVEC ACTES INTERNATIONAL (ACTI)
Langue
anglais
Auteur
FONSECA ; DUMAS ; ARMAND
Résumé / Abstract
Accurate gait event detection is a key factor for a reliable analysis of pathological gait. Detection of events using ground reaction forces is considered as the gold standard for detecting events. However, in general, only few steps are recorded over the force platforms during a trial. Therefore, several methods have been proposed to estimate gait events using analysis of marker trajectories. However, pathological gait is highly variable and no method has been yet consensually accepted as the best method for gait analysis. More recently, two methods were proposed using deep learning with high accuracy and flexibility requiring moderate computation time.
Source