A liveness detection method for face recognition based on optical flow field

By Wei Bao, Hong Li, Nan Li, Wei Jiang. International Conference on Image Analysis and Signal Processing, 2009

363 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/bec757b63093832ea47b14976f47e17abbadd8b5

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Biometrics Evaluation Under Spoofing Attacks (2014, methodology): “a) Methodology 1: In the first evaluation methodology, two decision threshold calculations are performed separately for the two scenarios [6], [28], [27], [7].”
Face Anti-spoofing: Visual Approach (2014, neutral): “In another work [34], Bao et al.”
Counter-measures to photo attacks in face recognition: A public database and a baseline (2011, methodology): “[15] the authors propose a method to detect attacks produced with planar media (such as paper or screens) using motion estimation by optical flow.”
Deep Tree Learning for Zero-Shot Face Anti-Spoofing (2019, background): “Works in [2, 5, 6, 18, 46] consider features in the temporal domain.”
Face anti-spoofing using patch and depth-based CNNs (2017, background): “They select Shearlet-based features to measure the image quality and the OFM of the face area as well as the whole scene area.”
Motion-based counter-measures to photo attacks in face recognition (2014, methodology): “Because RS2 makes even stronger assumptions about the motion characteristics for real-accesses, results for this system are even worse, peaking more than 75% HTER on Figure 9b.”
Biometric Antispoofing Methods: A Survey in Face Recognition (2014, background): “, nodding, smiling, looking in different directions) detected trough face and gaze tracking [134], [135] or through optical flow estimation [27], [103], [136], [137].”
Cross-Domain Face Presentation Attack Detection via Multi-Domain Disentangled Representation Learning (2020, background): “Besides, a number of methods also explored temporal cues of the whole face or individual face components for PAD, such as eye blink [41], mouth movement [23] and head rotation [3].”
Learn Convolutional Neural Network for Face Anti-Spoofing (2014, background): “More recently, Anjos et al. proposed to recognize spooﬁng attacks based on the correlation between optical ﬂows in foreground and background regions [1].”
Learning Deep Models for Face Anti-Spoofing: Binary or Auxiliary Supervision (2018, background): “Additionally, there are some works proposing temporal-speciﬁc features, e.g., Haralick features [4], motion mag [7], and optical ﬂow [6].”
Context based face anti-spoofing (2013, background): “video displays and photographs, and rigid 3D face masks differs significantly from that of real human faces which are complex non-rigid 3D objects [3, 7, 15, 27].”
Single-Side Domain Generalization for Face Anti-Spoofing (2020, methodology): “In contrast, the temporal-based methods leverage various temporal cues in consecutive face frames, such as rPPG [20, 21] and optical flow [2, 4].”
Face Recognition on Consumer Devices: Reflections on Replay Attacks (2015, background): “There has been a significant focus on solving the problem of presenting fake biometrics (spoofing attacks) to sensors [7]–[9], [11], [13], [33], [34].”
Face spoofing detection from single images using micro-texture analysis (2011, background): “In another work [2], Bao et al. also used optical flow for motion estimation for detecting attacks produced with planar media such as prints or screens.”
Deep Learning in the Wild (2018, neutral): “These techniques mainly focus on texture differences between real and spoofed images, differences in color space [7], Fourier spectra [30], or optical flow maps [6].”
Face liveness detection with component dependent descriptor (2013, background): “This response can be represented by biometric motions, such as eye blinking [23, 18], head rotation [2], mouth movement [13], and the holistic motion [12].”
Face Biometrics Under Spoofing Attacks: Vulnerabilities, Countermeasures, Open Issues, and Research Directions (2014, methodology): “In [5], Bao et al. proposed a method to detect attacks produced with planar media (such as paper or screens) using motion estimation by optical flow.”