• This code trains and implements liveness detection with video input

• Modified original model to Resnet50 to achieve better accuracy (Thanks Jason Chu!)

• Install steps:

Use Anaconda3

pip install imutils

pip install keras

pip install --upgrade h5py

pip install opencv-python

• Train & Deploy steps:

use ./gather.bat to get train image snapshots from video in ./videos folder. (images will be stored in ./dataset/fake & real)

run ./trainLiveness.bat to train images from ./dataset

run ./runLiveness.bat to see implementaion of Liveness Detection via video input (-c 0.5 is the threshold of "Fake" and "Liveness")

https://www.pyimagesearch.com/2019/03/11/liveness-detection-with-opencv

• Support it by clicking the ⭐ button on the upper right of this page. ✌️

• The example has been taken from pyimagesearch liveness-detection-with-opencv example and modified model to Resnet50

Copyright (C) 2020 Adrian Rosebrock, PyImageSearch, https://www.pyimagesearch.com/2019/03/11/liveness-detection-with-opencv/, accessed on March 11, 2019

Just make pull request. Thank you!