Face Liveness Detection Competition (LivDet-Face) - 2021

Purnapatra, Sandip; Smalt, Nic; Bahmani, Keivan; Das, Priyanka; Yambay, David; Mohammadi, Amir; George, Anjith; Bourlai, Thirimachos; Marcel, Sébastien; Schuckers, Stephanie; Fang, Meiling; Damer, Naser; Kuijper, Arjan; Kantarcı, Alperen; Demir, Basar; Yildiz, Zafer; Ghafoory, Zabi; Dertli, Hasan; Ekenel, Hazım Kemal; Vu, Son; Christophides, Vassilis; Liang, Dashuang; Zhang, Guanghao; Hao, Zhanlong; Liu, Junfu; Jin, Yong; Liu, Samo; Huang, Samuel H.; Kuei, Salieri; Singh, Jag Mohan; Ramachandra, Raghavendra

doi:10.1109/ijcb52358.2021.9484359

Cited by 25 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To increase samples, perturbed and/or combined versions of the samples are usually made from simple transformations such as rotation or noise addition. In this case, this type of strategy has been successfully used in liveness detection techniques in biometrics such as faces [75], iris [76] and even fingerprints [77].…”

Section: A Data Augmentationmentioning

confidence: 99%

A New Multi-Filter Framework for Texture Image Representation Improvement Using Set of Pattern Descriptors to Fingerprint Liveness Detection

Contreras

Nonato

Boaventura³

et al. 2022

IEEE Access

View full text Add to dashboard Cite

The use of user recognition and authentication systems has become very common and is part of everyday routines for many people, guaranteeing access to the automatic teller machines, entrance to the gym or even to smartphones. Among all the biometrics that can be analyzed in this type of system, the fingerprint is the most considered due to the ease of collection, the uniqueness of each user, and the large amount of solid theories and computational libraries available in the scientific literature. However, in recent years, the falsification of these biometrics with synthetic materials, known as spoofing, has become a real threat to these systems. To circumvent these effects without the addition of hardware devices, techniques based on the analysis of texture pattern descriptors were developed. In this work, we propose a new framework based on steps of data augmentation, image processing and replication, and feature fusion and reduction. The method has as main objective to improve the ability of classifiers, or sets of classifiers, to recognize life in fingerprints. Furthermore, it is proposed a generalization of vector representation of patterns described in matrix form from the systematic use of sets of mapping functions. All the proposed material was analyzed on the well-established benchmark of the Liveness Detection competition of the 2009, 2011, 2013 and 2015 editions, presenting an average accuracy of 97.77% and being a competitive strategy in relation to the other techniques that make up the state of the art of specialized literature.

show abstract

Section: A Data Augmentationmentioning

confidence: 99%

A New Multi-Filter Framework for Texture Image Representation Improvement Using Set of Pattern Descriptors to Fingerprint Liveness Detection

Contreras

Nonato

Boaventura³

et al. 2022

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Deep-learning based face PAD methods achieved a great progress in intra-dataset evaluations, however, the performance normally drops drastically when testing on unseen datasets [20]. This might be caused by the variations in the attacks and capture environments, such as illuminations and sensors.…”

Section: Multi-level Frequency Decomposition (Mfd)mentioning

confidence: 99%

Learnable Multi-level Frequency Decomposition and Hierarchical Attention Mechanism for Generalized Face Presentation Attack Detection

Fang¹,

Damer²,

Kirchbuchner³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

With the increased deployment of face recognition systems in our daily lives, face presentation attack detection (PAD) is attracting a lot of attention and playing a key role in securing face recognition systems. Despite the great performance achieved by the hand-crafted and deep learning based methods in intra-dataset evaluations, the performance drops when dealing with unseen scenarios. In this work, we propose a dual-stream convolution neural networks (CNNs) framework. One stream adapts four learnable frequency filters to learn features in the frequency domain, which are less influenced variations in sensors/illuminations. The other stream leverage the RGB images to complement the features of the frequency domain. Moreover, we propose a hierarchical attention module integration to join the information from the two streams at different stages by considering the nature of deep features in different layers of the CNN. The proposed method is evaluated in the intra-dataset and cross-dataset setups and the results demonstrates that our proposed approach enhances the generalizability in most experimental setups in comparison to state-of-the-art, including the methods designed explicitly for domain adaption/shift problem. We successfully prove the design of our proposed PAD solution in a stepwise ablation study that involves our proposed learnable frequency decomposition, our hierarchical attention module design, and the used loss function. Training codes and pre-trained models are publicly released 1 .

show abstract

“…For a reliable user experience, algorithms need to be trained on the target demographics across all application use-cases. This training is conducted on the entire (FR) pipeline: face detection [1,2], face identification [3,4] and presentation-attack-detection (PAD) [5][6][7]. PAD, in particular, introduces data complexity due to the adversarial nature of attack detection.…”

Section: Introductionmentioning

confidence: 99%

“…Monocular PAD methods are often based on fine artifact detection methods [6,7,13]. Spoofing articles are rarely 100% perfect representations of the real face, where artifacts can be observed in textural aberrations [14], facial "correctness" [15], and light reflectance [16].…”

Section: Introductionmentioning

confidence: 99%

Improving Monocular Facial Presentation–Attack–Detection Robustness with Synthetic Noise Augmentations

Hassani,

Diedrich,

Malik

2023

Sensors

View full text Add to dashboard Cite

We present a synthetic augmentation approach towards improving monocular face presentation–attack–detection (PAD) robustness to real-world noise additions. Face PAD algorithms secure authentication systems against spoofing attacks, such as pictures, videos, and 2D-inspired masks. Best-in-class PAD methods typically use 3D imagery, but these can be expensive. To reduce application cost, there is a growing field investigating monocular algorithms that detect facial artifacts. These approaches work well in laboratory conditions, but can be sensitive to the imaging environment (e.g., sensor noise, dynamic lighting, etc.). The ideal solution for noise robustness is training under all expected conditions; however, this is time consuming and expensive. Instead, we propose that physics-informed noise-augmentations can pragmatically achieve robustness. Our toolbox contains twelve sensor and lighting effect generators. We demonstrate that our toolbox generates more robust PAD features than popular augmentation methods in noisy test-evaluations. We also observe that the toolbox improves accuracy on clean test data, suggesting that it inherently helps discern spoof artifacts from imaging artifacts. We validate this hypothesis through an ablation study, where we remove liveliness pairs (e.g., live or spoof imagery only for participants) to identify how much real data can be replaced with synthetic augmentations. We demonstrate that using these noise augmentations allows us to achieve better test accuracy while only requiring 30% of participants to be fully imaged under all conditions. These findings indicate that synthetic noise augmentations are a great way to improve PAD, addressing noise robustness while simplifying data collection.

show abstract

Face Liveness Detection Competition (LivDet-Face) - 2021

Cited by 25 publications

References 12 publications

A New Multi-Filter Framework for Texture Image Representation Improvement Using Set of Pattern Descriptors to Fingerprint Liveness Detection

A New Multi-Filter Framework for Texture Image Representation Improvement Using Set of Pattern Descriptors to Fingerprint Liveness Detection

Learnable Multi-level Frequency Decomposition and Hierarchical Attention Mechanism for Generalized Face Presentation Attack Detection

Improving Monocular Facial Presentation–Attack–Detection Robustness with Synthetic Noise Augmentations

Contact Info

Product

Resources

About