Disguised Faces in the Wild 2019
01 Oct 2019-pp 0-0
TL;DR: The outcome of the Disguised Faces in the Wild 2019 competition is summarized in terms of the dataset used for evaluation, a brief review of the algorithms employed by the participants for this task, and the results obtained.
Abstract: Disguised face recognition has wide-spread applicability in scenarios such as law enforcement, surveillance, and access control. Disguise accessories such as sunglasses, masks, scarves, or make-up modify or occlude different facial regions which makes face recognition a challenging task. In order to understand and benchmark the state-of-the-art on face recognition in the presence of disguise variations, the Disguised Faces in the Wild 2019 (DFW2019) competition has been organized. This paper summarizes the outcome of the competition in terms of the dataset used for evaluation, a brief review of the algorithms employed by the participants for this task, and the results obtained. The DFW2019 dataset has been released with four evaluation protocols and baseline results obtained from two deep learning-based state-of-the-art face recognition models. The DFW2019 dataset has also been analyzed with respect to degrees of difficulty: (i) easy, (ii) medium, and (iii) hard. The dataset has been released as part of the International Workshop on Disguised Faces in the Wild at International Conference on Computer Vision (ICCV), 2019.
Citations
More filters
TL;DR: A comprehensive review of the recent developments on deep face recognition can be found in this paper, covering broad topics on algorithm designs, databases, protocols, and application scenes, as well as the technical challenges and several promising directions.
353 citations
TL;DR: Various applications and opportunities of SM multimodal data, latest advancements, current challenges, and future directions for the crisis informatics and other related research fields are highlighted.
Abstract: People increasingly use Social Media (SM) platforms such as Twitter and Facebook during disasters and emergencies to post situational updates including reports of injured or dead people, infrastructure damage, requests of urgent needs, and the like. Information on SM comes in many forms, such as textual messages, images, and videos. Several studies have shown the utility of SM information for disaster response and management, which encouraged humanitarian organizations to start incorporating SM data sources into their workflows. However, several challenges prevent these organizations from using SM data for response efforts. These challenges include near-real-time information processing, information overload, information extraction, summarization, and verification of both textual and visual content. We highlight various applications and opportunities of SM multimodal data, latest advancements, current challenges, and future directions for the crisis informatics and other related research fields.
69 citations
03 Apr 2020
TL;DR: Different ways in which the robustness of a face recognition algorithm is challenged, which can severely affect its intended working are summarized.
Abstract: Face recognition algorithms have demonstrated very high recognition performance, suggesting suitability for real world applications Despite the enhanced accuracies, robustness of these algorithms against attacks and bias has been challenged This paper summarizes different ways in which the robustness of a face recognition algorithm is challenged, which can severely affect its intended working Different types of attacks such as physical presentation attacks, disguise/makeup, digital adversarial attacks, and morphing/tampering using GANs have been discussed We also present a discussion on the effect of bias on face recognition models and showcase that factors such as age and gender variations affect the performance of modern algorithms The paper also presents the potential reasons for these challenges and some of the future research directions for increasing the robustness of face recognition models
53 citations
Cites background or methods from "Disguised Faces in the Wild 2019"
...…the top performing teams in the competition demonstrated high verification performance at higher False Acceptance Rates (Deng and Zafeririou 2019; Singh et al. 2019a), analysis of the submissions demonstrate low performance (less than 10% True Acceptance Rate) at 0% False Acceptance Rate; a…...
[...]
...In 2018, the Disguised Faces in the Wild (DFW) dataset (Singh et al. 2019b) was released as part of the International Workshop on DFW held in conjunction with CVPR2018....
[...]
...Recently, the DFW2019 competition (Singh et al. 2019a) has also contained a protocol for recognizing images under plastic surgery variations, where deep learning based baseline algorithms show around 50% verification accuracy at 0.01% False Acceptance Rate....
[...]
...Recently, the DFW2019 competition (Singh et al. 2019a) has also contained a protocol for recognizing images under plastic surgery variations, where deep learning based baseline algorithms show around 50% verification accuracy at 0....
[...]
...While the top performing teams in the competition demonstrated high verification performance at higher False Acceptance Rates (Deng and Zafeririou 2019; Singh et al. 2019a), analysis of the submissions demonstrate low performance (less than 10% True Acceptance Rate) at 0% False Acceptance Rate; a metric often used in stricter settings such as access control in highly secure locations....
[...]
01 Oct 2019
TL;DR: By using the authors' RetinaFace for face detection and alignment and Arc face for face feature embedding, this work achieves state-of-the-art performance on the DFW2019 challenge.
Abstract: Even though deep face recognition is extensively explored and remarkable advances have been achieved on large-scale in-the-wild dataset, disguised face recognition receives much less attention. Face feature embedding targeting on intra-class compactness and inter-class discrepancy is very challenging as high intra-class diversity and inter-class similarity are very common on the disguised face recognition dataset. In this report, we give the technical details of our submission to the DFW2019 challenge. By using our RetinaFace for face detection and alignment and ArcFace for face feature embedding, we achieve state-of-the-art performance on the DFW2019 challenge.
16 citations
Cites methods from "Disguised Faces in the Wild 2019"
...Method Impersonation Obfuscation Plastic Surgery Overall FAR 1e− 4 1e− 3 1e− 2 1e− 4 1e− 3 1e− 4 1e− 3 1e− 4 1e− 3 ResNet-50 [23] 38....
[...]
...On the Impersonation track, our solution is worse than the baseline method (LightCNN-29v2 [23]) provided by the organiser....
[...]
TL;DR: In this article, a comprehensive survey of works related to the topic of makeup presentation attack detection is provided, along with a critical discussion, and the vulnerability of a commercial off-the-shelf and an open-source face recognition system against makeup presentation attacks is assessed.
Abstract: The application of facial cosmetics may cause substantial alterations in the facial appearance, which can degrade the performance of facial biometrics systems. Additionally, it was recently demonstrated that makeup can be abused to launch so-called makeup presentation attacks. More precisely, an attacker might apply heavy makeup to obtain the facial appearance of a target subject with the aim of impersonation or to conceal their own identity. We provide a comprehensive survey of works related to the topic of makeup presentation attack detection, along with a critical discussion. Subsequently, we assess the vulnerability of a commercial off-the-shelf and an open-source face recognition system against makeup presentation attacks. Specifically, we focus on makeup presentation attacks with the aim of impersonation employing the publicly available Makeup Induced Face Spoofing (MIFS) and Disguised Faces in the Wild (DFW) databases. It is shown that makeup presentation attacks might seriously impact the security of face recognition systems. Further, we propose different image pair-based, i.e. differential, attack detection schemes which analyse differences in feature representations obtained from potential makeup presentation attacks and corresponding target face images. The proposed detection systems employ various types of feature extractors including texture descriptors, facial landmarks, and deep (face) representations. To distinguish makeup presentation attacks from genuine, i.e. bona fide presentations, machine learning-based classifiers are used. The classifiers are trained with a large number of synthetically generated makeup presentation attacks utilising a generative adversarial network for facial makeup transfer in conjunction with image warping. Experimental evaluations conducted using the MIFS database and a subset of the DFW database reveal that deep face representations achieve competitive detection equal error rates of 0.7% and 1.8%, respectively.
12 citations
References
More filters
08 Mar 2019
TL;DR: The disguised faces in the wild (DFW) dataset as discussed by the authors contains over 11,000 images of 1000 identities with variations across different types of disguise accessories, including impersonator and genuine obfuscated face images for each subject.
Abstract: Research in face recognition has seen tremendous growth over the past couple of decades. Beginning from algorithms capable of performing recognition in constrained environments, existing face recognition systems achieve very high accuracies on large-scale unconstrained face datasets. While upcoming algorithms continue to achieve improved performance, many of them are susceptible to reduced performance under disguise variations, one of the most challenging covariate of face recognition. In this paper, the disguised faces in the wild (DFW) dataset is presented, which contains over 11000 images of 1000 identities with variations across different types of disguise accessories (the DFW dataset link: http://iab-rubric.org/resources/dfw.html ). The images are collected from the Internet, resulting in unconstrained variations similar to real-world settings. This is a unique dataset that contains impersonator and genuine obfuscated face images for each subject. The DFW dataset has been analyzed in terms of three levels of difficulty: 1) easy; 2) medium; and 3) hard, in order to showcase the challenging nature of the problem. The dataset was released as part of the First International Workshop and Competition on DFW at the Conference on Computer Vision and Pattern Recognition, 2018. This paper presents the DFW dataset in detail, including the evaluation protocols, baseline results, performance analysis of the submissions received as part of the competition, and three levels of difficulties of the DFW challenge dataset.
52 citations
01 Feb 2016
TL;DR: A new database for face images under disguised and make-up appearances the development of face recognition algorithms under such covariates is presented and the experimental results suggest significant performance degradation in the capability of these matchers in automatically recognizing these faces.
Abstract: The accuracy of automated human face recognition algorithms can significantly degrade while recognizing same subjects under make-up and disguised appearances. Increasing constraints on enhanced security and surveillance requires enhanced accuracy from face recognition algorithms for faces under disguise and/or makeup. This paper presents a new database for face images under disguised and make-up appearances the development of face recognition algorithms under such covariates. This database has 2460 images from 410 different subjects and is acquired under real environment, focuses on make-up and disguises covariates and also provides ground truth (eye glass, goggle, mustache, beard) for every image. This can enable developed algorithms to automatically quantify their capability for identifying such important disguise attribute during the face recognition We also present comparative experimental results from two popular commercial matchers and from recent publications. Our experimental results suggest significant performance degradation in the capability of these matchers in automatically recognizing these faces. We also analyze face detection accuracy from these matchers. The experimental results underline the challenges in recognizing faces under these covariates. Availability of this new database in public domain will help to advance much needed research and development in recognizing make-up and disguised faces.
43 citations
"Disguised Faces in the Wild 2019" refers methods in this paper
...In 2016, the Disguise and Makeup dataset [19] was released, which contains disguised face images from publicly accessible websites....
[...]
01 Oct 2017
TL;DR: It is shown that Doppelganger mining, being inserted in the face representation learning process with joint prototype-based and exemplar-based supervision, significantly improves the discriminative power of learned face representations.
Abstract: In this paper we present Doppelganger mining - a method to learn better face representations. The main idea of this method is to maintain a list with the most similar identities for each identity in the training set. This list is used to generate better mini-batches by sampling pairs of similar-looking identities ("doppelgangers") together. It is especially useful for methods, based on exemplar-based supervision. Usually hard example mining comes with a price of necessity to use large mini-batches or substantial extra computation and memory cost, particularly for datasets with large numbers of identities. Our method needs only a negligible extra computation and memory. In our experiments on a benchmark dataset with 21,000 persons we show that Doppelganger mining, being inserted in the face representation learning process with joint prototype-based and exemplar-based supervision, significantly improves the discriminative power of learned face representations.
38 citations
"Disguised Faces in the Wild 2019" refers methods in this paper
...Fine-tuning is performed on the MS-Celeb-1M dataset [6] with Doppleganger Mining [17], Auxillary Embeddings [18], Embeddings Interpolations, and Priority Lists....
[...]
...The model is fine-tuned with Doppleganger Mining [17], Auxillary Embeddings [18], Embeddings Interpolations, and Priority Lists....
[...]
18 Jun 2018
TL;DR: The experiments on the challenging Disguised Faces in the Wild dataset show that hard example mining with auxiliary embeddings improves the discriminative power of learned representations.
Abstract: Hard example mining is an important part of the deep embedding learning. Most methods perform it at the mini-batch level. However, in the large-scale settings there is only a small chance that proper examples will appear in the same mini-batch and will be coupled into the hard example pairs or triplets. Doppelganger mining was previously proposed to increase this chance by means of class-wise similarity. This method ensures that examples of similar classes are sampled into the same mini-batch together. One of the drawbacks of this method is that it operates only at the class level, while there also might be a way to select appropriate examples within class in a more elaborated way than randomly. In this paper, we propose to use auxiliary embeddings for hard example mining. These embeddings are constructed in such way that similar examples have close embeddings in the cosine similarity sense. With the help of these embeddings it is possible to select new examples for the mini-batch based on their similarity with the already selected examples. We propose several ways to create auxiliary embeddings and use them to increase the number of potentially hard positive and negative examples in each mini-batch. Our experiments on the challenging Disguised Faces in the Wild (DFW) dataset show that hard example mining with auxiliary embeddings improves the discriminative power of learned representations.
38 citations
"Disguised Faces in the Wild 2019" refers methods in this paper
...Fine-tuning is performed on the MS-Celeb-1M dataset [6] with Doppleganger Mining [17], Auxillary Embeddings [18], Embeddings Interpolations, and Priority Lists....
[...]
...The model is fine-tuned with Doppleganger Mining [17], Auxillary Embeddings [18], Embeddings Interpolations, and Priority Lists....
[...]