Skip to content

Source Code for D-NetPAD (Densely Connected Network based Iris Presentation Attack Detection): https://arxiv.org/abs/2007.01381

Notifications You must be signed in to change notification settings

sharmaGIT/D-NetPAD

Repository files navigation

D-NetPAD

Code for Iris Presentation Attack Detection based on DenseNet Architecture.

Requirement

Pytorch, Numpy, Scipy, Pillow

Description

The D-NetPAD takes a cropped iris image as input and produces a PA score between 0 and 1, where 0 means bonafide and 1 means presentation attack. Sample cropped iris images are provided in CroppedImages folder.

Testing

The model can be downloaded from here. Copy the model into the Model folder and run the following command:

python test_D-NetPAD.py -imageFolder CroppedImages

PA score CSV file will be created in the folder of images.

Training

python train_D-NetPAD.py -csvPath cseFilePath -datasetPath datasetImagesPath -outputPath resultPath

CSV file contains ground truth of dataset images. The format of the dataset CSV file is

train,Live,imageFile1.png
train,Spoof,imageFile2.png
test,Live,imageFile3.png
test,Spoof,imageFile4.png

Fine Tuning

python fineTrain_D-NetPAD.py -csvPath cseFilePath -datasetPath datasetImagesPath -outputPath resultPath

Citation

If you are using the code, please cite the paper:

Renu Sharma, Arun Ross, D-NetPAD: An Explainable and Interpretable Iris Presentation Attack Detector, arXiv:2007.01381, 2020.

About

Source Code for D-NetPAD (Densely Connected Network based Iris Presentation Attack Detection): https://arxiv.org/abs/2007.01381

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages