deepevent 0.4

Deep Learning to identify gait events

DeepEvent Version 0.4

We propose a new application (DeepEvent) of long short term memory recurrent neural network to the automatic detection of gait events. The 3D position and velocity of the markers on the heel, toe and lateral malleolus were used by the network to determine Foot Strike (FS) and Foot Off (FO). The method was developed from 10526 FS and 9375 FO from 226 children. DeepEvent predicted FS within 5.5 ms and FO within 10.7 ms of the gold standard (automatic determination using force platform data) and was more accurate than common heuristic marker trajectory-based methods proposed in the literature and another deep learning method.

Requirement

Windows 64 bits only, python3.7, Keras, Tensorflow, pyBtk, Numpy, Scipy, GoogleDriveDownloader

All depedancies are available in pipy

Installation

pip install deepevent==0.4

Running DeepEvent

deepevent  -i FilenameIn.c3d -o FilenameOut.c3d
deepevent  --input FilenameIn.c3d --output FilenameOut.c3d
deepevent  --input FilenameIn.c3d

where FilenameIn.c3d is the c3d file to identify gait events, FilenameOut.c3d is the new file with the gait events.
In the last case, filenameIn.c3d is overwritten with gait events

Bibliography

Lempereur M., Rousseau F., Rémy-Néris O., Pons C., Houx L., Quellec G., Brochard S. (2019). A new deep learning-based method for the detection of gait events in children with gait disorders: Proof-of-concept and concurrent validity. Journal of Biomechanics. Volume 98, 2 January 2020, 109490.

Contacts

Original developer [e-mail]
Mathieu LEMPEREUR
CHRU de Brest - Hopital Morvan
Service MPR
2 avenue Foch
29609 BREST cedex
FRANCE

Forked developer [e-mail]
Fabien Leboeuf
Ingénieur "analyste du mouvement" du Pole 10, CHU Nantes, France
Chercheur associé de l'Université de Salford, Manchester, Royaume uni
Laboraratoire d'analyse du mouvement
85 rue saint Jacques
44093 Nantes
FRANCE