Face recognition algorithms, when used in the real world, can be very useful, but they can also be dangerous when biased toward certain demographics. So, it is essential to understand how these algorithms are trained and what factors affect their accuracy and fairness to build better ones. In this study, we shed some light on the effect of racial distribution in the training data on the performance of face recognition models. We conduct 16 different experiments with varying racial distributions of faces in the training data. We analyze these trained models using accuracy metrics, clustering metrics, UMAP projections, face quality, and decision thresholds. We show that a uniform distribution of races in the training datasets alone does not guarantee bias-free face recognition algorithms and how factors like face image quality play a crucial role. We also study the correlation between the clustering metrics and bias to understand whether clustering is a good indicator of bias. Finally, we introduce a metric called racial gradation to study the inter and intra race correlation in facial features and how they affect the learning ability of the face recognition models. With this study, we try to bring more understanding to an essential element of face recognition training, the data. A better understanding of the impact of training data on the bias of face recognition algorithms will aid in creating better datasets and, in turn, better face recognition systems.

Healthcare: According to the increasing size of Data, many CV applications are invented in the Healthcare industry. Security: By using face detection in the Surveillance cameras in various sectors, security firms and local authorities are using the advantages of CV applications. Autonomous Vehicles: Nowadays, autonomous vehicles are so popular. Starting from Tesla, many other companies try to develop their autonomous vehicle to keep up with the future technology. The logic behind these systems is also one of the CV applications, which continuously works around the car while it is self-driven.

As the use of deep learning in high impact domains becomes ubiquitous, it is increasingly important to assess the resilience of models. One such high impact domain is that of face recognition, with real world applications involving images affected by various degradations, such as motion blur or high exposure. Moreover, images captured across different attributes, such as gender and race, can also challenge the robustness of a face recognition algorithm. While traditional summary statistics suggest that the aggregate performance of face recognition models has continued to improve, these metrics do not directly measure the robustness or fairness of the models. Visual Psychophysics Sensitivity Analysis (VPSA) [1] provides a way to pinpoint the individual causes of failure by way of introducing incremental perturbations in the data. However, perturbations may affect subgroups differently. In this paper, we propose a new fairness evaluation based on robustness in the form of a generic framework that extends VPSA. With this framework, we can analyze the ability of a model to perform fairly for different subgroups of a population affected by perturbations, and pinpoint the exact failure modes for a subgroup by measuring targeted robustness. With the increasing focus on the fairness of models, we use face recognition as an example application of our framework and propose to compactly visualize the fairness analysis of a model via AUC matrices. We analyze the performance of common face recognition models and empirically show that certain subgroups are at a disadvantage when images are perturbed, thereby uncovering trends that were not visible using the model's performance on subgroups without perturbations.

In spite of the successful application in many fields, machine learning algorithms today suffer from notorious problems like vulnerability to adversarial examples. Beyond falling into the cat-and-mouse game between adversarial attack and defense, this paper provides alternative perspective to consider adversarial example and explore whether we can exploit it in benign applications. We first propose a novel taxonomy of visual information along task-relevance and semantic-orientation. The emergence of adversarial example is attributed to algorithm's utilization of task-relevant non-semantic information. While largely ignored in classical machine learning mechanisms, task-relevant non-semantic information enjoys three interesting characteristics as (1) exclusive to algorithm, (2) reflecting common weakness, and (3) utilizable as features. Inspired by this, we present brave new idea called benign adversarial attack to exploit adversarial examples for goodness in three directions: (1) adversarial Turing test, (2) rejecting malicious algorithm, and (3) adversarial data augmentation. Each direction is positioned with motivation elaboration, justification analysis and prototype applications to showcase its potential.

In a letter to congress sent on June 8th, IBM's CEO Arvind Krishna made a bold statement regarding the company's policy toward facial recognition. "IBM no longer offers general purpose IBM facial recognition or analysis software," says Krishna. "IBM firmly opposes and will not condone uses of any technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms, or any purpose which is not consistent with our values and Principles of Trust and Transparency." The company has halted all facial recognition development and disapproves or any technology that could lead to racial profiling. The ethics of face recognition technology have been in question for years. However, there has been little to no movement in the enactment of official laws barring the technology.

Almost 200 face recognition algorithms--a majority in the industry--had worse performance on nonwhite faces, according to a landmark study. What they tested: The US National Institute of Standards and Technology (NIST) tested every algorithm on two of the most common tasks for face recognition. The first, known as "one-to-one" matching, involves matching a photo of someone to another photo of the same person in a database. This is used to unlock smartphones or check passports, for example. The second, known as "one-to-many" searching, involves determining whether a photo of someone has any match in a database.

Artificial intelligence is a potentially world-changing technology. It could help cure cancers, control autonomous cars, and augment human intelligence. Or it could lead to a robot apocalypse and the downfall of humanity. It depends on who you ask. Artificial intelligence or AI simply means software used by computers to mimic aspects of human intelligence.

This fair usage agreement ("Agreement") is between CyberExtruder.com, Inc., a New York corporation with its principal office located at 1401 Valley Road, Wayne, New Jersey, 07470, USA ("Company") and the user of the Data Set, as defined below ("Licensee"). Whereas the Licensee is interested in the fair use of the Data Set for the non-commercial purposes of testing face recognition algorithms, and the Company wants to facilitate Licensee's testing of face recognition algorithms, the parties agree as follows: The Data Set contains 10,205 images of 1000 people collected randomly from the internet and is unrestricted with regard to the subject's pose, environmental lighting conditions, facial expression, subject's race and subject's age and contains images which are artistic impressions, drawings, paintings and other non-photographic representations of faces, and a multitude of facial occlusions like hats, glasses and makeup. All images are sized to 600 x 600 pixels and are stored with jpeg compression. LICENSE GRANT A non-exclusive, nontransferable, royalty-free license is granted to Licensee to use the Data Set on an appropriate computer system located at Licensee's premises. The Company is free, at its sole discretion, to distribute the Data Set to others and to use it for its own purposes.

Kairos snaps up EmotionReader, which can scan faces in a crowd and tell how audiences are reacting. EmotionReader, an Enterprise Ireland-backed facial recognition start-up, has been acquired by Miami-based Kairos in an undisclosed "multimillion-dollar" deal. Artificial intelligence (AI) then analyses viewer attention and emotional response, enabling media and brand owners to collect actionable insights and analytics for video. 'In our mission to fix biases in today's face recognition algorithms, we're thrilled to welcome to Kairos some of the best deep-learning talent in the world' – BRIAN BRACKEEN The company is the brainchild of Dr Padraig O'Leary and Dr Stephen Moore, and it was founded only last year. Moore, working from his Singapore base, is understood to have built an impressive R&D team in the south-east Asian country.

Deep neural network (DNN) architecture based models have high expressive power and learning capacity. However, they are essentially a black box method since it is not easy to mathematically formulate the functions that are learned within its many layers of representation. Realizing this, many researchers have started to design methods to exploit the drawbacks of deep learning based algorithms questioning their robustness and exposing their singularities. In this paper, we attempt to unravel three aspects related to the robustness of DNNs for face recognition: (i) assessing the impact of deep architectures for face recognition in terms of vulnerabilities to attacks inspired by commonly observed distortions in the real world that are well handled by shallow learning methods along with learning based adversaries; (ii) detecting the singularities by characterizing abnormal filter response behavior in the hidden layers of deep networks; and (iii) making corrections to the processing pipeline to alleviate the problem. Our experimental evaluation using multiple open-source DNN-based face recognition networks, including OpenFace and VGG-Face, and two publicly available databases (MEDS and PaSC) demonstrates that the performance of deep learning based face recognition algorithms can suffer greatly in the presence of such distortions. The proposed method is also compared with existing detection algorithms and the results show that it is able to detect the attacks with very high accuracy by suitably designing a classifier using the response of the hidden layers in the network. Finally, we present several effective countermeasures to mitigate the impact of adversarial attacks and improve the overall robustness of DNN-based face recognition.