Gesture_recognition_DVS

Prerequisites

1. Python 3.6
2. Tensorflow-1.2

Get the Dataset

The dataset can be obtained from the below link:

Link: https://pan.baidu.com/s/1FdohqlkF10wDhtDJ9KAxhQ , code: 7tmn

Ten people are involved to record gestures with only hand and arms in the field. There are 18 bin files recorded for one person. The background class is recorded in three illumination settings seperately each for 18 minutes.

Get the pretrained models

The trained models can be obtained from the below link:

Link: ^^^^^ , code: &&&

How to use the code

Prepare the data

1. Choose the denoise method.
2. Convert each bin files into images.

Training

1. Use train_*.py to train the networks for the proposed dataset.

Real-time testing

1. Use Celex_Realtime_*py to evaluate the trained networks.
2. After running the python file, copy the bin files to the directory(Celex_Camera)

Contact

For any question, please contact

[email protected]