PyTorchKR Papers

최신 논문

총 131편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2105.03790

Distribution Matching for Heterogeneous Multi-Task Learning: a Large-scale Face Study

Multi-Task Learning has emerged as a methodology in which multiple tasks are jointly learned by a shared learning algorithm, such as a DNN. MTL is based on the assumption that the tasks under consideration are related; therefore it exploits shared knowledge for improving performance on each individual task. Tasks are generally considered to be homogeneous, i.e., to refer to the same type of problem. Moreover, MTL is usually based on ground truth annotations with full, or partial overlap across tasks. In this work, we deal with heterogeneous MTL, simultaneously addressing detection, classification & regression problems. We explore task-relatedness as a means for co-training, in a weakly-supervised way, tasks that contain little, or even non-overlapping annotations. Task-relatedness is introduced in MTL, either explicitly through prior expert knowledge, or through data-driven studies. We propose a novel distribution matching approach, in which knowledge exchange is enabled between tasks, via matching of their predictions' distributions. Based on this approach, we build FaceBehaviorNet, the first framework for large-scale face analysis, by jointly learning all facial behavior tasks. We develop case studies for: i) continuous affect estimation, action unit detection, basic emotion recognition; ii) attribute detection, face identification. We illustrate that co-training via task relatedness alleviates negative transfer. Since FaceBehaviorNet learns features that encapsulate all aspects of facial behavior, we conduct zero-/few-shot learning to perform tasks beyond the ones that it has been trained for, such as compound emotion recognition. By conducting a very large experimental study, utilizing 10 databases, we illustrate that our approach outperforms, by large margins, the state-of-the-art in all tasks and in all databases, even in these which have not been used in its training.

arXiv:2104.09145

Face-GCN: A Graph Convolutional Network for 3D Dynamic Face Identification/Recognition

Face identification/recognition has significantly advanced over the past years. However, most of the proposed approaches rely on static RGB frames and on neutral facial expressions. This has two disadvantages. First, important facial shape cues are ignored. Second, facial deformations due to expressions can have an impact on the performance of such a method. In this paper, we propose a novel framework for dynamic 3D face identification/recognition based on facial keypoints. Each dynamic sequence of facial expressions is represented as a spatio-temporal graph, which is constructed using 3D facial landmarks. Each graph node contains local shape and texture features that are extracted from its neighborhood. For the classification/identification of faces, a Spatio-temporal Graph Convolutional Network (ST-GCN) is used. Finally, we evaluate our approach on a challenging dynamic 3D facial expression dataset.

arXiv:2104.03419

Towards On-Device Face Recognition in Body-worn Cameras

Face recognition technology related to recognizing identities is widely adopted in intelligence gathering, law enforcement, surveillance, and consumer applications. Recently, this technology has been ported to smartphones and body-worn cameras (BWC). Face recognition technology in body-worn cameras is used for surveillance, situational awareness, and keeping the officer safe. Only a handful of academic studies exist in face recognition using the body-worn camera. A recent study has assembled BWCFace facial image dataset acquired using a body-worn camera and evaluated the ResNet-50 model for face identification. However, for real-time inference in resource constraint body-worn cameras and privacy concerns involving facial images, on-device face recognition is required. To this end, this study evaluates lightweight MobileNet-V2, EfficientNet-B0, LightCNN-9 and LightCNN-29 models for face identification using body-worn camera. Experiments are performed on a publicly available BWCface dataset. The real-time inference is evaluated on three mobile devices. The comparative analysis is done with heavy-weight VGG-16 and ResNet-50 models along with six hand-crafted features to evaluate the trade-off between the performance and model size. Experimental results suggest the difference in maximum rank-1 accuracy of lightweight LightCNN-29 over best-performing ResNet-50 is \textbf{1.85\%} and the reduction in model parameters is \textbf{23.49M}. Most of the deep models obtained similar performances at rank-5 and rank-10. The inference time of LightCNNs is 2.1x faster than other models on mobile devices. The least performance difference of \textbf{14\%} is noted between LightCNN-29 and Local Phase Quantization (LPQ) descriptor at rank-1. In most of the experimental settings, lightweight LightCNN models offered the best trade-off between accuracy and the model size in comparison to most of the models.

arXiv:2103.17107공식 코드 · Papers with Code

Facial expression and attributes recognition based on multi-task learning of lightweight neural networks

In this paper, we examine the multi-task training of lightweight convolutional neural networks for face identification and classification of facial attributes (age, gender, ethnicity) trained on cropped faces without margins. It is shown that it is still necessary to fine-tune these networks in order to predict facial expressions. Several models are presented based on MobileNet, EfficientNet and RexNet architectures. It was experimentally demonstrated that our models are characterized by the state-of-the-art emotion classification accuracy on AffectNet dataset and near state-of-the-art results in age, gender and race recognition for UTKFace dataset. Moreover, it is shown that the usage of our neural network as a feature extractor of facial regions in video frames and concatenation of several statistical functions (mean, max, etc.) leads to 4.5% higher accuracy than the previously known state-of-the-art single models for AFEW and VGAF datasets from the EmotiW challenges.

arXiv:1805.05682

Task-dependent modulation of the visual sensory thalamus assists visual-speech recognition

The cerebral cortex modulates early sensory processing via feed-back connections to sensory pathway nuclei. The functions of this top-down modulation for human behavior are poorly understood. Here, we show that top-down modulation of the visual sensory thalamus (the lateral geniculate body, LGN) is involved in visual-speech recognition. In two independent functional magnetic resonance imaging (fMRI) studies, LGN response increased when participants processed fast-varying features of articulatory movements required for visual-speech recognition, as compared to temporally more stable features required for face identification with the same stimulus material. The LGN response during the visual-speech task correlated positively with the visual-speech recognition scores across participants. In addition, the task-dependent modulation was present for speech movements and did not occur for control conditions involving non-speech biological movements. In face-to-face communication, visual speech recognition is used to enhance or even enable understanding what is said. Speech recognition is commonly explained in frameworks focusing on cerebral cortex areas. Our findings suggest that task-dependent modulation at subcortical sensory stages has an important role for communication: Together with similar findings in the auditory modality the findings imply that task-dependent modulation of the sensory thalami is a general mechanism to optimize speech recognition.

arXiv:2011.13360

ClusterFace: Joint Clustering and Classification for Set-Based Face Recognition

Deep learning technology has enabled successful modeling of complex facial features when high quality images are available. Nonetheless, accurate modeling and recognition of human faces in real world scenarios `on the wild' or under adverse conditions remains an open problem. When unconstrained faces are mapped into deep features, variations such as illumination, pose, occlusion, etc., can create inconsistencies in the resultant feature space. Hence, deriving conclusions based on direct associations could lead to degraded performance. This rises the requirement for a basic feature space analysis prior to face recognition. This paper devises a joint clustering and classification scheme which learns deep face associations in an easy-to-hard way. Our method is based on hierarchical clustering where the early iterations tend to preserve high reliability. The rationale of our method is that a reliable clustering result can provide insights on the distribution of the feature space, that can guide the classification that follows. Experimental evaluations on three tasks, face verification, face identification and rank-order search, demonstrates better or competitive performance compared to the state-of-the-art, on all three experiments.

arXiv:2010.08391

Reconstructing A Large Scale 3D Face Dataset for Deep 3D Face Identification

Deep learning methods have brought many breakthroughs to computer vision, especially in 2D face recognition. However, the bottleneck of deep learning based 3D face recognition is that it is difficult to collect millions of 3D faces, whether for industry or academia. In view of this situation, there are many methods to generate more 3D faces from existing 3D faces through 3D face data augmentation, which are used to train deep 3D face recognition models. However, to the best of our knowledge, there is no method to generate 3D faces from 2D face images for training deep 3D face recognition models. This letter focuses on the role of reconstructed 3D facial surfaces in 3D face identification and proposes a framework of 2D-aided deep 3D face identification. In particular, we propose to reconstruct millions of 3D face scans from a large scale 2D face database (i.e.VGGFace2), using a deep learning based 3D face reconstruction method (i.e.ExpNet). Then, we adopt a two-phase training approach: In the first phase, we use millions of face images to pre-train the deep convolutional neural network (DCNN), and in the second phase, we use normal component images (NCI) of reconstructed 3D face scans to train the DCNN. Extensive experimental results illustrate that the proposed approach can greatly improve the rank-1 score of 3D face identification on the FRGC v2.0, the Bosphorus, and the BU-3DFE 3D face databases, compared to the model trained by 2D face images. Finally, our proposed approach achieves state-of-the-art rank-1 scores on the FRGC v2.0 (97.6%), Bosphorus (98.4%), and BU-3DFE (98.8%) databases. The experimental results show that the reconstructed 3D facial surfaces are useful and our 2D-aided deep 3D face identification framework is meaningful, facing the scarcity of 3D faces.

arXiv:2010.05222공식 코드 · Papers with Code

Partial FC: Training 10 Million Identities on a Single Machine

Face recognition has been an active and vital topic among computer vision community for a long time. Previous researches mainly focus on loss functions used for facial feature extraction network, among which the improvements of softmax-based loss functions greatly promote the performance of face recognition. However, the contradiction between the drastically increasing number of face identities and the shortage of GPU memories is gradually becoming irreconcilable. In this paper, we thoroughly analyze the optimization goal of softmax-based loss functions and the difficulty of training massive identities. We find that the importance of negative classes in softmax function in face representation learning is not as high as we previously thought. The experiment demonstrates no loss of accuracy when training with only 10\% randomly sampled classes for the softmax-based loss functions, compared with training with full classes using state-of-the-art models on mainstream benchmarks. We also implement a very efficient distributed sampling algorithm, taking into account model accuracy and training efficiency, which uses only eight NVIDIA RTX2080Ti to complete classification tasks with tens of millions of identities. The code of this paper has been made available https://github.com/deepinsight/insightface/tree/master/recognition/partial_fc.

arXiv:2010.00975

Taking Modality-free Human Identification as Zero-shot Learning

Human identification is an important topic in event detection, person tracking, and public security. There have been numerous methods proposed for human identification, such as face identification, person re-identification, and gait identification. Typically, existing methods predominantly classify a queried image to a specific identity in an image gallery set (I2I). This is seriously limited for the scenario where only a textual description of the query or an attribute gallery set is available in a wide range of video surveillance applications (A2I or I2A). However, very few efforts have been devoted towards modality-free identification, i.e., identifying a query in a gallery set in a scalable way. In this work, we take an initial attempt, and formulate such a novel Modality-Free Human Identification (named MFHI) task as a generic zero-shot learning model in a scalable way. Meanwhile, it is capable of bridging the visual and semantic modalities by learning a discriminative prototype of each identity. In addition, the semantics-guided spatial attention is enforced on visual modality to obtain representations with both high global category-level and local attribute-level discrimination. Finally, we design and conduct an extensive group of experiments on two common challenging identification tasks, including face identification and person re-identification, demonstrating that our method outperforms a wide variety of state-of-the-art methods on modality-free human identification.

arXiv:2009.11458

BWCFace: Open-set Face Recognition using Body-worn Camera

With computer vision reaching an inflection point in the past decade, face recognition technology has become pervasive in policing, intelligence gathering, and consumer applications. Recently, face recognition technology has been deployed on bodyworn cameras to keep officers safe, enabling situational awareness and providing evidence for trial. However, limited academic research has been conducted on this topic using traditional techniques on datasets with small sample size. This paper aims to bridge the gap in the state-of-the-art face recognition using bodyworn cameras (BWC). To this aim, the contribution of this work is two-fold: (1) collection of a dataset called BWCFace consisting of a total of 178K facial images of 132 subjects captured using the body-worn camera in in-door and daylight conditions, and (2) open-set evaluation of the latest deep-learning-based Convolutional Neural Network (CNN) architectures combined with five different loss functions for face identification, on the collected dataset. Experimental results on our BWCFace dataset suggest a maximum of 33.89% Rank-1 accuracy obtained when facial features are extracted using SENet-50 trained on a large scale VGGFace2 facial image dataset. However, performance improved up to a maximum of 99.00% Rank-1 accuracy when pretrained CNN models are fine-tuned on a subset of identities in our BWCFace dataset. Equivalent performances were obtained across body-worn camera sensor models used in existing face datasets. The collected BWCFace dataset and the pretrained/ fine-tuned algorithms are publicly available to promote further research and development in this area. A downloadable link of this dataset and the algorithms is available by contacting the authors.

arXiv:2009.04776

Self-supervised Depth Denoising Using Lower- and Higher-quality RGB-D sensors

Consumer-level depth cameras and depth sensors embedded in mobile devices enable numerous applications, such as AR games and face identification. However, the quality of the captured depth is sometimes insufficient for 3D reconstruction, tracking and other computer vision tasks. In this paper, we propose a self-supervised depth denoising approach to denoise and refine depth coming from a low quality sensor. We record simultaneous RGB-D sequences with unzynchronized lower- and higher-quality cameras and solve a challenging problem of aligning sequences both temporally and spatially. We then learn a deep neural network to denoise the lower-quality depth using the matched higher-quality data as a source of supervision signal. We experimentally validate our method against state-of-the-art filtering-based and deep denoising techniques and show its application for 3D object reconstruction tasks where our approach leads to more detailed fused surfaces and better tracking.

arXiv:2008.12262

DeepFake Detection Based on the Discrepancy Between the Face and its Context

We propose a method for detecting face swapping and other identity manipulations in single images. Face swapping methods, such as DeepFake, manipulate the face region, aiming to adjust the face to the appearance of its context, while leaving the context unchanged. We show that this modus operandi produces discrepancies between the two regions. These discrepancies offer exploitable telltale signs of manipulation. Our approach involves two networks: (i) a face identification network that considers the face region bounded by a tight semantic segmentation, and (ii) a context recognition network that considers the face context (e.g., hair, ears, neck). We describe a method which uses the recognition signals from our two networks to detect such discrepancies, providing a complementary detection signal that improves conventional real vs. fake classifiers commonly used for detecting fake images. Our method achieves state of the art results on the FaceForensics++, Celeb-DF-v2, and DFDC benchmarks for face manipulation detection, and even generalizes to detect fakes produced by unseen methods.

arXiv:2008.06674

BroadFace: Looking at Tens of Thousands of People at Once for Face Recognition

The datasets of face recognition contain an enormous number of identities and instances. However, conventional methods have difficulty in reflecting the entire distribution of the datasets because a mini-batch of small size contains only a small portion of all identities. To overcome this difficulty, we propose a novel method called BroadFace, which is a learning process to consider a massive set of identities, comprehensively. In BroadFace, a linear classifier learns optimal decision boundaries among identities from a large number of embedding vectors accumulated over past iterations. By referring more instances at once, the optimality of the classifier is naturally increased on the entire datasets. Thus, the encoder is also globally optimized by referring the weight matrix of the classifier. Moreover, we propose a novel compensation method to increase the number of referenced instances in the training stage. BroadFace can be easily applied on many existing methods to accelerate a learning process and obtain a significant improvement in accuracy without extra computational burden at inference stage. We perform extensive ablation studies and experiments on various datasets to show the effectiveness of BroadFace, and also empirically prove the validity of our compensation method. BroadFace achieves the state-of-the-art results with significant improvements on nine datasets in 1:1 face verification and 1:N face identification tasks, and is also effective in image retrieval.

arXiv:2008.04469공식 코드 · Papers with Code

Key-Nets: Optical Transformation Convolutional Networks for Privacy Preserving Vision Sensors

Modern cameras are not designed with computer vision or machine learning as the target application. There is a need for a new class of vision sensors that are privacy preserving by design, that do not leak private information and collect only the information necessary for a target machine learning task. In this paper, we introduce key-nets, which are convolutional networks paired with a custom vision sensor which applies an optical/analog transform such that the key-net can perform exact encrypted inference on this transformed image, but the image is not interpretable by a human or any other key-net. We provide five sufficient conditions for an optical transformation suitable for a key-net, and show that generalized stochastic matrices (e.g. scale, bias and fractional pixel shuffling) satisfy these conditions. We motivate the key-net by showing that without it there is a utility/privacy tradeoff for a network fine-tuned directly on optically transformed images for face identification and object detection. Finally, we show that a key-net is equivalent to homomorphic encryption using a Hill cipher, with an upper bound on memory and runtime that scales quadratically with a user specified privacy parameter. Therefore, the key-net is the first practical, efficient and privacy preserving vision sensor based on optical homomorphic encryption.

arXiv:2008.04821

Unified Representation Learning for Cross Model Compatibility

We propose a unified representation learning framework to address the Cross Model Compatibility (CMC) problem in the context of visual search applications. Cross compatibility between different embedding models enables the visual search systems to correctly recognize and retrieve identities without re-encoding user images, which are usually not available due to privacy concerns. While there are existing approaches to address CMC in face identification, they fail to work in a more challenging setting where the distributions of embedding models shift drastically. The proposed solution improves CMC performance by introducing a light-weight Residual Bottleneck Transformation (RBT) module and a new training scheme to optimize the embedding spaces. Extensive experiments demonstrate that our proposed solution outperforms previous approaches by a large margin for various challenging visual search scenarios of face recognition and person re-identification.

arXiv:2008.03353

Hybrid Score- and Rank-level Fusion for Person Identification using Face and ECG Data

Uni-modal identification systems are vulnerable to errors in sensor data collection and are therefore more likely to misidentify subjects. For instance, relying on data solely from an RGB face camera can cause problems in poorly lit environments or if subjects do not face the camera. Other identification methods such as electrocardiograms (ECG) have issues with improper lead connections to the skin. Errors in identification are minimized through the fusion of information gathered from both of these models. This paper proposes a methodology for combining the identification results of face and ECG data using Part A of the BioVid Heat Pain Database containing synchronized RGB-video and ECG data on 87 subjects. Using 10-fold cross-validation, face identification was 98.8% accurate, while the ECG identification was 96.1% accurate. By using a fusion approach the identification accuracy improved to 99.8%. Our proposed methodology allows for identification accuracies to be significantly improved by using disparate face and ECG models that have non-overlapping modalities.

arXiv:2007.11987

Multi-Metric Evaluation of Thermal-to-Visual Face Recognition

In this paper, we aim to address the problem of heterogeneous or cross-spectral face recognition using machine learning to synthesize visual spectrum face from infrared images. The synthesis of visual-band face images allows for more optimal extraction of facial features to be used for face identification and/or verification. We explore the ability to use Generative Adversarial Networks (GANs) for face image synthesis, and examine the performance of these images using pre-trained Convolutional Neural Networks (CNNs). The features extracted using CNNs are applied in face identification and verification. We explore the performance in terms of acceptance rate when using various similarity measures for face verification.

arXiv:2007.06402

Nested Learning For Multi-Granular Tasks

Standard deep neural networks (DNNs) are commonly trained in an end-to-end fashion for specific tasks such as object recognition, face identification, or character recognition, among many examples. This specificity often leads to overconfident models that generalize poorly to samples that are not from the original training distribution. Moreover, such standard DNNs do not allow to leverage information from heterogeneously annotated training data, where for example, labels may be provided with different levels of granularity. Furthermore, DNNs do not produce results with simultaneous different levels of confidence for different levels of detail, they are most commonly an all or nothing approach. To address these challenges, we introduce the concept of nested learning: how to obtain a hierarchical representation of the input such that a coarse label can be extracted first, and sequentially refine this representation, if the sample permits, to obtain successively refined predictions, all of them with the corresponding confidence. We explicitly enforce this behavior by creating a sequence of nested information bottlenecks. Looking at the problem of nested learning from an information theory perspective, we design a network topology with two important properties. First, a sequence of low dimensional (nested) feature embeddings are enforced. Then we show how the explicit combination of nested outputs can improve both the robustness and the accuracy of finer predictions. Experimental results on Cifar-10, Cifar-100, MNIST, Fashion-MNIST, Dbpedia, and Plantvillage demonstrate that nested learning outperforms the same network trained in the standard end-to-end fashion.

arXiv:2006.13026공식 코드 · Papers with Code

Deep Polynomial Neural Networks

Deep Convolutional Neural Networks (DCNNs) are currently the method of choice both for generative, as well as for discriminative learning in computer vision and machine learning. The success of DCNNs can be attributed to the careful selection of their building blocks (e.g., residual blocks, rectifiers, sophisticated normalization schemes, to mention but a few). In this paper, we propose $\Pi$-Nets, a new class of function approximators based on polynomial expansions. $\Pi$-Nets are polynomial neural networks, i.e., the output is a high-order polynomial of the input. The unknown parameters, which are naturally represented by high-order tensors, are estimated through a collective tensor factorization with factors sharing. We introduce three tensor decompositions that significantly reduce the number of parameters and show how they can be efficiently implemented by hierarchical neural networks. We empirically demonstrate that $\Pi$-Nets are very expressive and they even produce good results without the use of non-linear activation functions in a large battery of tasks and signals, i.e., images, graphs, and audio. When used in conjunction with activation functions, $\Pi$-Nets produce state-of-the-art results in three challenging tasks, i.e. image generation, face verification and 3D mesh representation learning. The source code is available at \url{https://github.com/grigorisg9gr/polynomial_nets}.

arXiv:2006.01615

A Multi-Task Comparator Framework for Kinship Verification

Approaches for kinship verification often rely on cosine distances between face identification features. However, due to gender bias inherent in these features, it is hard to reliably predict whether two opposite-gender pairs are related. Instead of fine tuning the feature extractor network on kinship verification, we propose a comparator network to cope with this bias. After concatenating both features, cascaded local expert networks extract the information most relevant for their corresponding kinship relation. We demonstrate that our framework is robust against this gender bias and achieves comparable results on two tracks of the RFIW Challenge 2020. Moreover, we show how our framework can be further extended to handle partially known or unknown kinship relations.