skeleton-AEs

An implementation of the paper Estimating skeleton-based gait abnormality index by sparse deep auto-encoder

Requirements

• Python
• Numpy
• TensorFlow
• Scikit-learn

Notice

• The code was implemented to directly work on DIRO gait dataset
• Please download the skeleton data and put the npz file into the folder dataset

Usage

python3 main.py

Example of output

Default training and test sets

(9, 9, 1200, 75)
Finish loading data
(6000, 17)
(4800, 17)
(38400, 17)

Training X...
length 1: AUC = 0.817
Training Y...
length 1: AUC = 0.763
Training Z...
length 1: AUC = 0.619

simple sum:
length 1: AUC = 0.748
length 20: AUC = 0.812
length 1200: AUC = 0.836

weighted sum:
length 1: AUC = 0.856
length 20: AUC = 0.911
length 1200: AUC = 0.945