Anomaly Detection-Based Unknown Face Presentation Attack Detection

Baweja, Yashasvi; Oza, Poojan; Perera, Pramuditha; Patel, Vishal M.

doi:10.1109/ijcb48548.2020.9304935

Cited by 28 publications

(16 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This feature is passed through the open-set classifier via path (1) to evaluate the cross entropy loss L cls . Then, the image corresponding to the obtained latent feature is generated by passing the feature through the decoder following path (2). The decoded image is used to calculate its difference to the corresponding clean image based on the reconstruction loss L rec .…”

Section: Proposed Methodsmentioning

confidence: 99%

“…A significant improvement has been achieved in the image classification task since the advent of deep convolutional neural networks (CNNs) [14]. The promising performance in classification has contributed to many real-world computer vision applications [41,37,40,38,39,36,42,45,47,20,2,46]. However, This work was conducted prior to joining AWS AI Labs when the author was affiliated with Johns Hopkins University.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Open-set Adversarial Defense

Shao

Perera

Yuen

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Open-set recognition and adversarial defense study two key aspects of deep learning that are vital for real-world deployment. The objective of open-set recognition is to identify samples from open-set classes during testing, while adversarial defense aims to defend the network against images with imperceptible adversarial perturbations. In this paper, we show that open-set recognition systems are vulnerable to adversarial attacks. Furthermore, we show that adversarial defense mechanisms trained on known classes do not generalize well to open-set samples. Motivated by this observation, we emphasize the need of an Open-Set Adversarial Defense (OSAD) mechanism. This paper proposes an Open-Set Defense Network (OSDN) as a solution to the OSAD problem. The proposed network uses an encoder with feature-denoising layers coupled with a classifier to learn a noise-free latent feature representation. Two techniques are employed to obtain an informative latent feature space with the objective of improving open-set performance. First, a decoder is used to ensure that clean images can be reconstructed from the obtained latent features. Then, self-supervision is used to ensure that the latent features are informative enough to carry out an auxiliary task. We introduce a testing protocol to evaluate OSAD performance and show the effectiveness of the proposed method in multiple object classification datasets. The implementation code of the proposed method is available at: https://github.com/rshaojimmy/ECCV2020-OSAD.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Open-set Adversarial Defense

Shao

Perera

Yuen

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Within the published works, it is possible to find reinforcement learning approaches [3], 3D-CNNs [19], a two stages approach relying on blinking [8] and several other colour-based methods [20,2]. The background usage is addressed in some works [27,1,22,30,16], however, they did not perform comparative studies regarding performance, with and without the background, of several approaches. It is possible to find this comparison in other works, however, the proposed methods are based on conventional machine learning instead of end-to-end deep learning [31,18].…”

Section: Related Workmentioning

confidence: 99%

Myope Models -- Are face presentation attack detection models short-sighted?

Neto¹,

Sequeira²,

Cardoso³

2021

Preprint

View full text Add to dashboard Cite

Presentation attacks are recurrent threats to biometric systems, where impostors attempt to bypass these systems. Humans often use background information as contextual cues for their visual system. Yet, regarding face-based systems, the background is often discarded, since face presentation attack detection (PAD) models are mostly trained with face crops. This work presents a comparative study of face PAD models (including multi-task learning, adversarial training and dynamic frame selection) in two settings: with and without crops. The results show that the performance is consistently better when the background is present in the images. The proposed multi-task methodology beats the state-of-the-art results on the ROSE-Youtu dataset by a large margin with an equal error rate of 0.2%. Furthermore, we analyze the models' predictions with Grad-CAM++ with the aim to investigate to what extent the models focus on background elements that are known to be useful for human inspection. From this analysis we can conclude that the background cues are not relevant across all the attacks. Thus, showing the capability of the model to leverage the background information only when necessary.

show abstract

“…FAS can be also formulated as a domain adaptation/generalization problem in [28,34,60], where disentangled learning [41,75], adversarial learning [56], and meta learning [10,37,38,57,61] are adopt to improve the model's generalization capacity on unseen scenarios. Besides, several anomaly detection [1,17,36,46], zero-shot [40,51] and continuous learning [49,53] approaches are proposed for unknown PAs detection. Though the generalization capacity has been improved by those techniques, they are still restricted to labeled data in source domains.…”

Section: Related Workmentioning

confidence: 99%

Consistency Regularization for Deep Face Anti-Spoofing

Wang¹,

Yu²,

Wang³

et al. 2021

Preprint

View full text Add to dashboard Cite

Face anti-spoofing (FAS) plays a crucial role in securing face recognition systems. Empirically, given an image, a model with more consistent output on different views of this image usually performs better, as shown in Fig. 1. Motivated by this exciting observation, we conjecture that encouraging feature consistency of different views may be a promising way to boost FAS models.In this paper, we explore this way thoroughly by enhancing both Embedding-level and Prediction-level Consistency Regularization (EPCR) in FAS. Specifically, at the embedding-level, we design a dense similarity loss to maximize the similarities between all positions of two intermediate feature maps in a self-supervised fashion; while at the prediction-level, we optimize the mean square error between the predictions of two views. Notably, our EPCR is free of annotations and can directly integrate into semi-supervised learning schemes. Considering different application scenarios, we further design five diverse semisupervised protocols to measure semi-supervised FAS techniques. We conduct extensive experiments to show that EPCR can significantly improve the performance of several supervised and semi-supervised tasks on benchmark datasets. The codes and protocols will be released at https://github.com/clks-wzz/EPCR.

show abstract

Anomaly Detection-Based Unknown Face Presentation Attack Detection

Cited by 28 publications

References 32 publications

Open-set Adversarial Defense

Open-set Adversarial Defense

Myope Models -- Are face presentation attack detection models short-sighted?

Consistency Regularization for Deep Face Anti-Spoofing

Contact Info

Product

Resources

About