Week 1:
1.Reading 3 paper about CNN.
4. Setting up a programming environment, using Ubuntu to produce a appropriate Python ennvironment.
Week 2:
1. Reading 3 paper about opencv and dlib.
Week 3:
1. Reading 1 paper about CNN's working theory.
2. Starting coding easy detecting program to lock the portrait.
Week 4:
1. We determine the outline of the report, and start wirting introduction.
Week 5:
1. Wirting the introduction.
2. Starting building our own dataset based on opencv and webcam.
Week 6:
1. Wirting the related work.
2. Building the model of liveness detection, and training this model.
Week 7:
1. Choosing 3 people to shoot videos, including printed photos and repaly. And cut it with fit photo, then make it as fake dataset. Downloading real face dataset in web. According to this 2 dataset to train the net.
Week 8:
1. Trying to train more net with more dataset, and trying to find the best net work.
Week 9:
1. Building a best vgg network, which can detect real and fake precisely.
Week 10:
1. Fixing some bugs.
Week 11:
1. Writing the preliminary repoter.
Week 12:
1. Starting adding one more convolution and testing the effect.
Week 13:
1. Compared previous network and new network. Impoving the new network
Week 14:
1. Finding some problem with new network, and trying to fix it.
Week 15:
1. Writing the repoter.
Week 16:
1. Writing the repoter.
Week 17:
1. Making a poster.
Week 18:
1. Sloving any problem with repoter and poster.