Face Anti-Spoofing based on image-quality measures

To edit a toolchain, please use a modern browser (Mozilla Firefox 3.0+, Google Chrome 1+, Apple Safari 3+, Opera 9.5+, Microsoft Internet Explorer 9+)

This toolchain implements a face anti-spoofing algorithm based on image-quality measures (IQM) and LDA.

The algorithm expects video files as input. A two-class classifier is constructed using a set of IQM. The real accesses are considered as the positive, and the attacks as the negative class.

The toolchain consists of the following steps:

  1. Feature extraction. This step extracts IQM features from the frames of the input video.
  2. Training. This step trains a binary classifier via LDA, on the training data.
  3. Classification. This step classifies the videos in the development-set and the test-set.
  4. Evaluation. This step calculates error rates and plots different performance curves.

The toolchain is based upon the work presented in [Chingovska12]. All the algorithms presented there can be readily implemented and used with this toolchain.

[Chingovska12]
  1. Chingovska, A. Anjos, S. Marcel: On the effectiveness of local binary patterns in face anti-spoofing. BIOSIG 2012
Updated Name Databases/Protocols Analyzers
sbhatta/sbhatta/iqm-face-antispoofing-test/2/replay2-antispoofing-iqm-lda replay/2@grandtest sbhatta/iqm_spoof_eer_analyzer/9
Terms of Service | Contact Information | BEAT platform version 1.2.3 | © Idiap Research Institute - 2013-2017