Face Anti Spoofing using Patch based CNN
The main objective of the project is to find a solution for one of the serious problems of the Facial recognition. We use MS-USSA dataset (Copyrighted under the name of Michigan State University) for the training of the dataset. For the project,we use Patch based approach. The faces in the images are identified using Haar Cascades. A new CNN is created and the faces extracted are given as training features and the model is trained.
For any queries regarding dataset, please visit the following link http://biometrics.cse.msu.edu/Publications/Databases/MSU_USSA/