Deep face recognition: A survey
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.
About: This article is published in Neurocomputing.The article was published on 2021-03-14 and is currently open access. It has received 353 citations till now. The article focuses on the topics: Deep learning & Feature extraction.
Citations
More filters
18 Nov 2019
TL;DR: This paper scientifically defines the three key problems of achieving cooperative intelligence, which are cooperative perception, cooperative decision and cooperative learning and proposes a system architecture and components design of cooperative intelligence system for autonomous driving.
Abstract: With the rapid development of deep learning, artificial intelligence (AI) has been widely used in many fields and gradually replaced a part of human jobs. However, the approach of improving intelligent capability of single agent is not enough to achieve complicated tasks in ever-changing environments. Cooperative intelligence (CI) is regarded as a promising way to solve this problem. In this paper, we scientifically define the three key problems of achieving cooperative intelligence, which are cooperative perception, cooperative decision and cooperative learning. We illustrate each problem with a scenario of autonomous driving as well as a brief survey of related research works. Meanwhile, we propose a system architecture and components design of cooperative intelligence system for autonomous driving.
2 citations
26 Jun 2021
TL;DR: In this article, the authors used evolutionary neural networks with uncertainty estimation to estimate a person's age from a facial image, which is a challenging problem with clinical applications, such as predicting the patient's appearance and perceived age.
Abstract: Estimating a person's age from a facial image is a challenging problem with clinical applications. Several medical aesthetics treatments have been developed that alter the skin texture and other facial features, with the goal of potentially improving patient's appearance and perceived age. In this paper, this effect was evaluated using evolutionary neural networks with uncertainty estimation. First, a realistic dataset was obtained from clinical studies that makes it possible to estimate age more reliably than e.g. datasets of celebrity images. Second, a neuroevolution approach was developed that customizes the architecture, learning, and data augmentation hyperparameters and the loss function to this task. Using state-of-the-art computer vision architectures as a starting point, evolution improved their original accuracy significantly, eventually outperforming the best human optimizations in this task. Third, the reliability of the age predictions was estimated using RIO, a Gaussian-Process-based uncertainty model. Evaluation on a real-world Botox treatment dataset shows that the treatment has a quantifiable result: The patients' estimated age is reduced significantly compared to placebo treatments. The study thus shows how AI can be harnessed in a new role: To provide an objective quantitative measure of a subjective perception, in this case the proposed effectiveness of medical aesthetics treatments.
2 citations
05 Sep 2021
TL;DR: In this paper, the authors implemented a system to monitor student attendance at South Ural State University based on automatic recognition of students' faces using neural networks, which allows to easily add a new person to the system without additional training of neural networks within the system.
Abstract: In this research we implement a system to monitor student attendance at South Ural State University. Our system is based on the automatic recognition of students' faces using neural networks. In this paper, we investigate the process, methods and algorithms of face recognition to build a student attendance control system that allows to easily add a new person to the system without additional training of neural networks within the system. In working, we identified and studied the parameters that reduce the efficiency of the recognition, as well as equipping our system with algorithms that reduce the negative impact of these parameters. We conducted experiments that prove the effectiveness of our solution.
2 citations
DOI•
15 Sep 2021TL;DR: In this paper, an ensemble of deep CNN networks was used for face recognition using different feature selection methods and three types of classifiers: support vector machine, a random forest of decision trees, and softmax built into the CNN classifier.
Abstract: The paper considers the problem of recognition of face images using an ensemble of deep CNN networks. The solution combines different feature selection methods and three types of classifiers: support vector machine, a random forest of decision trees, and softmax built into the CNN classifier. Deep learning fulfills an important role in the developed system. The numerical descriptors created in the last locally connected convolutional layer of CNN flattened to the form of a vector, are subject to four different selection mechanisms. Their results are delivered to the three classifiers which are the members of the ensemble. The developed system was tested on the problem of face recognition. The dataset was composed of 68 classes of greyscale images. The results of experiments have shown significant improvement of class recognition resulted from the application of the ensemble.
2 citations
TL;DR: In this article , the analysis of the main Spanish political candidates for the elections to be held on April 2019 is presented, focusing on the Facial Expression Analysis (FEA), a technique widely used in neuromarketing research.
Abstract: This article presents the analysis of the main Spanish political candidates for the elections to be held on April 2019. The analysis focuses on the Facial Expression Analysis (FEA), a technique widely used in neuromarketing research. It allows to identify the micro-expressions that are very brief, involuntary. They are signals of hidden emotions that cannot be controlled voluntarily. The video with the final interventions of every candidate has been post-processed using the classification algorithms given by the iMotions's AFFDEX platform. We have then analyzed these data. Firstly, we have identified and compare the basic emotions showed by each politician. Second, we have associated the basic emotions with specific moments of the candidate's speech, identifying the topics they address and relating them directly to the expressed emotion. Third, we have analyzed whether the differences shown by each candidate in every emotion are statistically significant. In this sense, we have applied the non-parametric chi-squared goodness-of-fit test. We have also considered the ANOVA analysis in order to test whether, on average, there are differences between the candidates. Finally, we have checked if there is consistency between the results provided by different surveys from the main media in Spain regarding the evaluation of the debate and those obtained in our empirical analysis. A predominance of negative emotions has been observed. Some inconsistencies were found between the emotion expressed in the facial expression and the verbal content of the message. Also, evidences got from statistical analysis confirm that the differences observed between the various candidates with respect to the basic emotions, on average, are statistically significant. In this sense, this article provides a methodological contribution to the analysis of the public figures' communication, which could help politicians to improve the effectiveness of their messages identifying and evaluating the intensity of the expressed emotions.
2 citations
References
More filters
27 Jun 2016
TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.
123,388 citations
Proceedings Article•
03 Dec 2012TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Abstract: We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overriding in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.
73,978 citations
Proceedings Article•
04 Sep 2014TL;DR: This work investigates the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting using an architecture with very small convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers.
Abstract: In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.
55,235 citations
07 Jun 2015
TL;DR: Inception as mentioned in this paper is a deep convolutional neural network architecture that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14).
Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.
40,257 citations
08 Dec 2014
TL;DR: A new framework for estimating generative models via an adversarial process, in which two models are simultaneously train: a generative model G that captures the data distribution and a discriminative model D that estimates the probability that a sample came from the training data rather than G.
Abstract: We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to ½ everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the generated samples.
38,211 citations