discriminate the CG from PG

For the purpose of finishing my project of graduation, I'm gonna accomplish this projcet. So the final goal of this project is to discriminate the CG(computer graphics) from PG(photographic images) with Artificial Intelligence, literally, when this program acquire new inputs which is out of range of training dataset, it's gonna classify the inputs belong to either CG subset or PG subset, more precisely, it will make a steganalysis on each input graphics to determine whethter or not it contains artificial patchs.

1. Cut into patchs (650 X 650)
2. Ongoing confidential part...
3. CNN based model
4. Comparative experiments
   1. H comparsion
   2. Quality factor comparsion