UVAD Database Access in Bob¶
This package provides an interface to the Unicamp Video-Attack Database (UVAD) database. The original data files need to be downloaded separately. After you have downloaded the dataset, you need to configure bob.db.uvad to find the dataset:
$ bob config set bob.db.uvad.directory /path/to/downloaded/dataset
If you use this database, please cite the following publication:
@ARTICLE{7017526,
author={Pinto, A. and Robson Schwartz, W. and Pedrini, H. and De Rezende Rocha, A.},
journal={Information Forensics and Security, IEEE Transactions on},
title={Using Visual Rhythms for Detecting Video-Based Facial Spoof Attacks},
year={2015},
month={May},
volume={10},
number={5},
pages={1025-1038},
keywords={Authentication;Biometrics (access control);Databases;Face;Feature extraction;Histograms;Noise;Unicamp Video-Attack Database;Video-based Face Spoofing;Video-based face spoofing;Visual Rhythm, Video-based Attacks;impersonation detection in facial biometric systems;unicamp video-attack database;video-based attacks;visual rhythm},
doi={10.1109/TIFS.2015.2395139},
ISSN={1556-6013},}
Package Documentation¶
-
bob.db.uvad.
UVAD_FRAME_SHAPE
= (3, 720, 1024)¶ Shape of the video frames in the uvad database.
-
class
bob.db.uvad.
Database
(original_directory=None, name='uvad', pad_file_class=None, original_extension=None, **kwargs)¶ Bases:
bob.pad.base.database.FileListPadDatabase
The database interface for the OULU-NPU dataset.
-
annotations
(padfile)[source]¶ Reads the annotations for the given padfile.
Parameters: padfile ( File
) – The file object for which the annotations should be read.Returns: The annotations as a dictionary, e.g.: {'0': {'reye':(re_y,re_x), 'leye':(le_y,le_x)}, ...}
Return type: dict
-
frame_shape
¶
-
-
class
bob.db.uvad.
File
(attack_type, client_id, path, file_id=None)¶ Bases:
bob.pad.face.database.VideoPadFile
The file objects of the OULU-NPU dataset.
-
annotations
¶
-
frame_shape
¶ Returns the size of each frame in this database.
Returns: The (#Channels, Height, Width) which is UVAD_FRAME_SHAPE
.Return type: (int, int, int)
-
frames
¶ Yields the frames of the biofile one by one.
Yields: numpy.array
– A frame of the video. The size is (3, 1920, 1080).
-