Deep Spatial Gradient and Temporal Depth Learning for Face Anti-Spoofing

@article{Wang2020DeepSG,
  title={Deep Spatial Gradient and Temporal Depth Learning for Face Anti-Spoofing},
  author={Zezheng Wang and Zitong Yu and Chenxu Zhao and Xiangyu Zhu and Yunxiao Qin and Qiusheng Zhou and Feng Zhou and Zhen Lei},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2020},
  pages={5041-5050}
}
  • Zezheng Wang, Zitong Yu, Zhen Lei
  • Published 18 March 2020
  • Computer Science
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Face anti-spoofing is critical to the security of face recognition systems. Depth supervised learning has been proven as one of the most effective methods for face anti-spoofing. Despite the great success, most previous works still formulate the problem as a single-frame multi-task one by simply augmenting the loss with depth, while neglecting the detailed fine-grained information and the interplay between facial depths and moving patterns. In contrast, we design a new approach to detect… 
View on IEEE
View With Semantic Reader
53 Citations
Highly Influential Citations
3
Background Citations
27
Methods Citations
25
Results Citations
1

53 Citations

Citation Type

Citation Type

Face Anti-Spoofing with Human Material Perception
TLDR
The Bilateral Convolutional Networks (BCN) is proposed, which is able to capture intrinsic material-based patterns via aggregating multi-level bilateral macro- and micro- information and Multi-level Feature Refinement Module (MFRM) and multi-head supervision are utilized to learn more robust features.
Analyzing Facial Temporal Patterns for Face Anti-Spoofing
TLDR
Two temporal modelling approaches to face anti-spoofing tasks are proposed to analyze the temporal patterns of mid-level facial attributes in spectral domain, aiming to find the unique frequency patterns of real face and each attack, respectively.
Revisiting Pixel-Wise Supervision for Face Anti-Spoofing
TLDR
A novel pyramid supervision is proposed, which guides deep models to learn both local details and global semantics from multi-scale spatial context in face anti-spoofing, to improve the performance beyond existing pixel-wise supervision frameworks and enhance the model’s interpretability.
Multi-Perspective Features Learning for Face Anti-Spoofing
TLDR
A Multi-Perspective Feature Learning Network (MPFLN) is proposed to extract representative features from the perspectives of Real + Mask vs. Photo + Video and Real + Video vs. Real vs. Spoof to alleviate the decline in the discrimination of extracted features and achieve comparable performance with state-of-the-art methods.
Deep Learning for Face Anti-Spoofing: A Survey
TLDR
This paper presents the first comprehensive review of recent advances in deep learning based FAS and collects and analyzes the latest methods specially designed for domain generalization and open-set FAS.
Attended-Auxiliary Supervision Representation for Face Anti-spoofing
TLDR
An attendedauxiliary supervision approach for radical exploitation is presented, which automatically concentrates on the most important regions of the input, that is, those that make significant contributions towards distinguishing the spoof cases from live faces, leading to notable improvements in performance.
Multi-Modal Face Anti-Spoofing Based on Central Difference Networks
TLDR
The central difference convolutional networks (CDCN) is extended to a multimodal version, intending to capture intrinsic spoofing patterns among three modalities (RGB, depth and infrared), and an elaborate study about singlemodal based CDCN is given.
Dual Spoof Disentanglement Generation for Face Anti-spoofing with Depth Uncertainty Learning
TLDR
A lightweight Depth Uncertainty Module (DUM) is introduced, which alleviates the adverse effects of noisy samples by depth uncertainty learning and can be flexibly integrated with any depth supervised network for face anti-spoofing.
Adversarial learning and decomposition-based domain generalization for face anti-spoofing
Adaptive Normalized Representation Learning for Generalizable Face Anti-Spoofing
TLDR
A novel perspective of face anti-spoofing that focuses on the normalization selection in the feature extraction process is proposed, which adaptively selects feature normalization methods according to the inputs, aiming to learn domain-agnostic and discriminative representation.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 57 REFERENCES
Face Liveness Detection by rPPG Features and Contextual Patch-Based CNN
TLDR
A generalized method exploiting both rPPG and texture features for face anti-spoofing task by exploiting multi-scale long-term statistical spectral features with variant granularities and a contextual patch-based convolutional neural network.
Exploring Hypergraph Representation On Face Anti-Spoofing Beyond 2d Attacks
TLDR
This paper addresses 3D face anti-spoofing via the proposed Hypergraph Convolutional Neural Networks (HGCNN) via a computation-efficient and posture-invariant face representation with only a few key points on hypergraphs.
Learning Deep Models for Face Anti-Spoofing: Binary or Auxiliary Supervision
TLDR
This paper argues the importance of auxiliary supervision to guide the learning toward discriminative and generalizable cues, and introduces a new face anti-spoofing database that covers a large range of illumination, subject, and pose variations.
Face Anti-Spoofing: Model Matters, so Does Data
  • Xiao Yang, Wenhan Luo, Wei Liu
  • Computer Science
    2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2019
TLDR
A data collection solution along with a data synthesis technique to simulate digital medium-based face spoofing attacks, and a novel Spatio-Temporal Anti-Spoof Network (STASN) that can distinguish spoof faces by extracting features from a variety of regions to seek out subtle evidences.
Face anti-spoofing using patch and depth-based CNNs
TLDR
A novel two-stream CNN-based approach for face anti-spoofing is proposed, by extracting the local features and holistic depth maps from the face images, which facilitate CNN to discriminate the spoof patches independent of the spatial face areas.
A Dataset and Benchmark for Large-Scale Multi-Modal Face Anti-Spoofing
TLDR
A large-scale multi-modal dataset, namely CASIA-SURF, is introduced, which is the largest publicly available dataset for face anti-spoofing in terms of both subjects and visual modalities and a new multi- modal fusion method is presented, which performs feature re-weighting to select the more informative channel features while suppressing the less useful ones for each modal.
An original face anti-spoofing approach using partial convolutional neural network
TLDR
This work extracts the deep partial features from the convolutional neural network (CNN) to distinguish the real and fake faces and uses the block principle component analysis (PCA) method to reduce the dimensionality of features that can avoid the over-fitting problem.
Improving Face Anti-Spoofing by 3D Virtual Synthesis
TLDR
A method to synthesize virtual spoof data in 3D space to alleviate the problem of expensive spoof data acquisition and open up new possibilities for advancing face anti-spoofing using cheap and large-scale synthetic data.
Learn Convolutional Neural Network for Face Anti-Spoofing
TLDR
Instead of designing feature by ourselves, the deep convolutional neural network is relied on to learn features of high discriminative ability in a supervised manner and combined with some data pre-processing, the face anti-spoofing performance improves drastically.
3D Mask Face Anti-spoofing with Remote Photoplethysmography
TLDR
A novel approach to 3D mask face anti-spoofing from a new perspective is proposed, by analysing heartbeat signal through remote Photoplethysmography (rPPG) to extract discriminative local heartbeat signal patterns so that an imposter can better be detected regardless of the material and quality of the mask.
...
1
2
3
4
5
...