Existing public face datasets are strongly biased toward Caucasian faces, and other races (e.g., Latino) are significantly underrepresented. This can lead to inconsistent model accuracy, limit the applicability of face analytic systems to non-White race groups, and adversely affect research findings based on such skewed data. To mitigate the race bias in these datasets, we construct a novel face image dataset, containing 108,501 images, with an emphasis of balanced race composition in the dataset. We define 7 race groups: White, Black, Indian, East Asian, Southeast Asian, Middle East, and Latino. Images were collected from the YFCC-100M Flickr dataset and labeled with race, gender, and age groups. Evaluations were performed on existing face attribute datasets as well as novel image datasets to measure generalization performance. We find that the model trained from our dataset is substantially more accurate on novel datasets and the accuracy is consistent between race and gender groups.
Rutgers University（罗格斯大学）；HPI / University of Potsdam（哈索·普拉特纳研究所/波茨坦大学）
Recent breakthroughs in self supervised training have led to a new class of pretrained vision language models. While there have been investigations of bias in multimodal models, they have mostly focused on gender and racial bias, giving much less attention to other relevant groups, such as minorities with regard to religion, nationality, sexual orientation, or disabilities. This is mainly due to lack of suitable benchmarks for such groups. We seek to address this gap by providing a visual and textual bias benchmark called MMBias, consisting of around 3,800 images and phrases covering 14 popula…
In this paper, we introduce FairFaceGAN, a fairness-aware facial Image-to-Image translation model, mitigating the problem of unwanted translation in protected attributes (e.g., gender, age, race) during facial attributes editing. Unlike existing models, FairFaceGAN learns fair representations with two separate latents - one related to the target attributes to translate, and the other unrelated to them. This strategy enables FairFaceGAN to separate the information about protected attributes and that of target attributes. It also prevents unwanted translation in protected attributes while target…
Facial forgery by deepfakes has raised severe societal concerns. Several solutions have been proposed by the vision community to effectively combat the misinformation on the internet via automated deepfake detection systems. Recent studies have demonstrated that facial analysis-based deep learning models can discriminate based on protected attributes. For the commercial adoption and massive roll-out of the deepfake detection technology, it is vital to evaluate and understand the fairness (the absence of any prejudice or favoritism) of deepfake detectors across demographic variations such as ge…
Recent research has questioned the fairness of face-based recognition and attribute classification methods (such as gender and race) for dark-skinned people and women. Ocular biometrics in the visible spectrum is an alternate solution over face biometrics, thanks to its accuracy, security, robustness against facial expression, and ease of use in mobile devices. With the recent COVID-19 crisis, ocular biometrics has a further advantage over face biometrics in the presence of a mask. However, fairness of ocular biometrics has not been studied till now. This first study aims to explore the fairne…
Automated gender classification has important applications in many domains, such as demographic research, law enforcement, online advertising, as well as human-computer interaction. Recent research has questioned the fairness of this technology across gender and race. Specifically, the majority of the studies raised the concern of higher error rates of the face-based gender classification system for darker-skinned people like African-American and for women. However, to date, the majority of existing studies were limited to African-American and Caucasian only. The aim of this paper is to invest…
Published studies have suggested the bias of automated face-based gender classification algorithms across gender-race groups. Specifically, unequal accuracy rates were obtained for women and dark-skinned people. To mitigate the bias of gender classifiers, the vision community has developed several strategies. However, the efficacy of these mitigation strategies is demonstrated for a limited number of races mostly, Caucasian and African-American. Further, these strategies often offer a trade-off between bias and classification accuracy. To further advance the state-of-the-art, we leverage the p…
Recent research has explored the possibility of automatically deducing information such as gender, age and race of an individual from their biometric data. While the face modality has been extensively studied in this regard, the iris modality less so. In this paper, we first review the medical literature to establish a biological basis for extracting gender and race cues from the iris. Then, we demonstrate that it is possible to use simple texture descriptors, like BSIF (Binarized Statistical Image Feature) and LBP (Local Binary Patterns), to extract gender and race attributes from an NIR ocul…
The U.S. Department of Homeland Security, The Maryland Test Facility Science and Technology Directorate（美国国土安全部、马里兰州测试设施科技总监办公室）
Human face features can be used to determine individual identity as well as demographic information like gender and race. However, the extent to which black-box commercial face recognition algorithms (CFRAs) use gender and race features to determine identity is poorly understood despite increasing deployments by government and industry. In this study, we quantified the degree to which gender and race features influenced face recognition similarity scores between different people, i.e. non-mated scores. We ran this study using five different CFRAs and a sample of 333 diverse test subjects. As a…
We demonstrate an approach to face attribute detection that retains or improves attribute detection accuracy across gender and race subgroups by learning demographic information prior to learning the attribute detection task. The system, which we call InclusiveFaceNet, detects face attributes by transferring race and gender representations learned from a held-out dataset of public race and gender identities. Leveraging learned demographic representations while withholding demographic inference from the downstream face attribute detection task preserves potential users' demographic privacy whil…
Czech Technical University in Prague（捷克布拉格技术大学）；Universitat Oberta de Catalunya（加泰罗尼亚开放大学）；Computer Vision Center（计算机视觉中心）；Universitat de Barcelona（巴塞罗那大学）；The Queen’s University of Belfast（贝尔法斯特女王大学）；Anyvision
This work summarizes the 2020 ChaLearn Looking at People Fair Face Recognition and Analysis Challenge and provides a description of the top-winning solutions and analysis of the results. The aim of the challenge was to evaluate accuracy and bias in gender and skin colour of submitted algorithms on the task of 1:1 face verification in the presence of other confounding attributes. Participants were evaluated using an in-the-wild dataset based on reannotated IJB-C, further enriched by 12.5K new images and additional labels. The dataset is not balanced, which simulates a real world scenario where …
Ask Me Again Differently: GRAS for Measuring Bias in Vision Language Models on Gender, Race, Age, and Skin Tone
Guru Gobind Singh Indraprastha University（古鲁·戈宾德·辛格印度教普拉斯塔大学）；AI Institute, University of South Carolina（南卡罗来纳大学人工智能研究所）；University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Indian Institute of Technology Patna, India（印度理工学院帕纳布分校）
Faces form the basis for a rich variety of judgments in humans, yet the underlying features remain poorly understood. Although fine-grained distinctions within a race might more strongly constrain possible facial features used by humans than in case of coarse categories such as race or gender, such fine grained distinctions are relatively less studied. Fine-grained race classification is also interesting because even humans may not be perfectly accurate on these tasks. This allows us to compare errors made by humans and machines, in contrast to standard object detection tasks where human perfo…
A biased dataset is a dataset that generally has attributes with an uneven class distribution. These biases have the tendency to propagate to the models that train on them, often leading to a poor performance in the minority class. In this project, we will explore the extent to which various data augmentation methods alleviate intrinsic biases within the dataset. We will apply several augmentation techniques on a sample of the UTKFace dataset, such as undersampling, geometric transformations, variational autoencoders (VAEs), and generative adversarial networks (GANs). We then trained a classif…
Michigan State University（密歇根州立大学）；University of Wisconsin – Madison（威斯康星大学麦迪逊分校）
Recent research has established the possibility of deducing soft-biometric attributes such as age, gender and race from an individual's face image with high accuracy. However, this raises privacy concerns, especially when face images collected for biometric recognition purposes are used for attribute analysis without the person's consent. To address this problem, we develop a technique for imparting soft biometric privacy to face images via an image perturbation methodology. The image perturbation is undertaken using a GAN-based Semi-Adversarial Network (SAN) - referred to as PrivacyNet - that…
University of Maryland（马里兰大学）；University of Chicago（芝加哥大学）
Many existing works have made great strides towards reducing racial bias in face recognition. However, most of these methods attempt to rectify bias that manifests in models during training instead of directly addressing a major source of the bias, the dataset itself. Exceptions to this are BUPT-Balancedface/RFW and Fairface, but these works assume that primarily training on a single race or not racially balancing the dataset are inherently disadvantageous. We demonstrate that these assumptions are not necessarily valid. In our experiments, training on only African faces induced less bias than…
Recent studies have demonstrated that deep learning models can discriminate based on protected classes like race and gender. In this work, we evaluate bias present in deepfake datasets and detection models across protected subgroups. Using facial datasets balanced by race and gender, we examine three popular deepfake detectors and find large disparities in predictive performances across races, with up to 10.7% difference in error rate between subgroups. A closer look reveals that the widely used FaceForensics++ dataset is overwhelmingly composed of Caucasian subjects, with the majority being f…
Fairness Without Labels: Pseudo-Balancing for Bias Mitigation in Face Gender Classification
National Institute of Informatics, Japan（日本信息机构国家研究所）；INESC-ID, Instituto Superior Técnico, University of Lisbon, Portugal（里斯本大学技术高等学院、INESC-ID，葡萄牙）；University of Amsterdam, The Netherlands（荷兰阿姆斯特丹大学）；LIMSI, CNRS / Université Paris-Saclay, France（法国CNRS/巴黎萨克雷大学LIMSI）
CUNY City College（纽约市立大学城市学院）；Graduate Center（研究生中心）
We propose a deep learning-based feature fusion approach for facial computing including face recognition as well as gender, race and age detection. Instead of training a single classifier on face images to classify them based on the features of the person whose face appears in the image, we first train four different classifiers for classifying face images based on race, age, gender and identification (ID). Multi-task features are then extracted from the trained models and cross-task-feature training is conducted which shows the value of fusing multimodal features extracted from multi-tasks. W…
Automated computer vision systems have been applied in many domains including security, law enforcement, and personal devices, but recent reports suggest that these systems may produce biased results, discriminating against people in certain demographic groups. Diagnosing and understanding the underlying true causes of model biases, however, are challenging tasks because modern computer vision systems rely on complex black-box models whose behaviors are hard to decode. We propose to use an encoder-decoder network developed for image attribute manipulation to synthesize facial images varying in…
In recent times, there have been increasing accusations on artificial intelligence systems and algorithms of computer vision of possessing implicit biases. Even though these conversations are more prevalent now and systems are improving by performing extensive testing and broadening their horizon, biases still do exist. One such class of systems where bias is said to exist is facial recognition systems, where bias has been observed on the basis of gender, ethnicity, skin tone and other facial attributes. This is even more disturbing, given the fact that these systems are used in practically ev…
We evaluate the state-of-the-art multimodal "visual semantic" model CLIP ("Contrastive Language Image Pretraining") for biases related to the marking of age, gender, and race or ethnicity. Given the option to label an image as "a photo of a person" or to select a label denoting race or ethnicity, CLIP chooses the "person" label 47.9% of the time for White individuals, compared with 5.0% or less for individuals who are Black, East Asian, Southeast Asian, Indian, or Latino or Hispanic. The model is more likely to rank the unmarked "person" label higher than labels denoting gender for Male indivi…
Large scale image classification models trained on top of popular datasets such as Imagenet have shown to have a distributional skew which leads to disparities in prediction accuracies across different subsections of population demographics. A lot of approaches have been made to solve for this distributional skew using methods that alter the model pre, post and during training. We investigate one such approach - which uses a multi-label softmax loss with cross-entropy as the loss function instead of a binary cross-entropy on a multi-label classification problem on the Inclusive Images dataset …
The growth in electronic transactions and human machine interactions rely on the information such as gender, age, expression and ethnicity provided by the face image. In order to obtain these information, feature extraction plays a major role. In this paper, retrieval of age, gender, expression and race information from an individual face image is analysed using different feature extraction methods. The performance of four major feature extraction methods such as Active Appearance Model (AAM), Gabor wavelets, Local Binary Pattern (LBP) and Wavelet Decomposition (WD) are analyzed for gender rec…
Similarity scores in face recognition represent the proximity between pairs of images as computed by a matching algorithm. Given a large set of images and the proximities between all pairs, a similarity score space is defined. Cluster analysis was applied to the similarity score space to develop various taxonomies. Given the number of subjects in the dataset, we used hierarchical methods to aggregate images of the same subject. We also explored the hierarchy above and below the subject level, including clusters that reflect gender and ethnicity. Evidence supports the existence of clustering by…
Vision-and-language tasks have increasingly drawn more attention as a means to evaluate human-like reasoning in machine learning models. A popular task in the field is visual question answering (VQA), which aims to answer questions about images. However, VQA models have been shown to exploit language bias by learning the statistical correlations between questions and answers without looking into the image content: e.g., questions about the color of a banana are answered with yellow, even if the banana in the image is green. If societal bias (e.g., sexism, racism, ableism, etc.) is present in t…
Computer vision models learn to perform a task by capturing relevant statistics from training data. It has been shown that models learn spurious age, gender, and race correlations when trained for seemingly unrelated tasks like activity recognition or image captioning. Various mitigation techniques have been presented to prevent models from utilizing or learning such biases. However, there has been little systematic comparison between these techniques. We design a simple but surprisingly effective visual recognition benchmark for studying bias mitigation. Using this benchmark, we provide a tho…
