Filters








132,787 Hits in 5.4 sec

Toward End-to-End Face Recognition Through Alignment Learning

Yuanyi Zhong, Jiansheng Chen, Bo Huang
2017 IEEE Signal Processing Letters  
In this paper we study the possibility of alignment learning in end-to-end face recognition, in which neither prior knowledge on facial landmarks nor artificially defined geometric transformations are  ...  In most systems, the face alignment module is implemented independently. This has actually caused difficulties in the designing and training of end-to-end face recognition models.  ...  Through an end-to-end learning, the face alignment and the facial feature extraction may interact with each other so as to achieve a joint optimum in term of the recognition task.  ... 
doi:10.1109/lsp.2017.2715076 fatcat:rbea5eqebjfh7ih66kt37abkoe

The Elements of End-to-end Deep Face Recognition: A Survey of Recent Advances [article]

Hang Du, Hailin Shi, Dan Zeng, Xiao-Ping Zhang, Tao Mei
2021 arXiv   pre-print
To start with, we present an overview of the end-to-end deep face recognition.  ...  Given a natural image or video frame as input, an end-to-end deep face recognition system outputs the face feature for recognition.  ...  Then, we give a detailed discussion about the future work towards end-to-end deep face recognition. 6.2.1 Face detection.  ... 
arXiv:2009.13290v4 fatcat:vlconzbbyzee5g3s7xnbjgv3ey

Towards the Design of an End-to-End Automated System for Image and Video-based Recognition [article]

Rama Chellappa, Jun-Cheng Chen, Rajeev Ranjan, Swami Sankaranarayanan, Amit Kumar, Vishal M. Patel, Carlos D. Castillo
2016 arXiv   pre-print
We then present the design details of a deep learning system for end-to-end unconstrained face verification/recognition.  ...  Over many decades, researchers working in object recognition have longed for an end-to-end automated system that will simply accept 2D or 3D image or videos as inputs and output the labels of objects in  ...  present a case study in designing an end-to-end face verification/recognition system using deep learning networks.  ... 
arXiv:1601.07883v1 fatcat:f6iphm7tyzg5vav2ju7tkfmu44

Towards the design of an end-to-end automated system for image and video-based recognition

Rama Chellappa, Jun-Cheng Chen, Rajeev Ranjan, Swami Sankaranarayanan, Amit Kumar, Vishal M. Patel, Carlos D. Castillo
2016 2016 Information Theory and Applications Workshop (ITA)  
We then present the design details of a deep learning system for endto-end unconstrained face verification/recognition.  ...  Over many decades, researchers working in object recognition have longed for an end-to-end automated system that will simply accept 2D or 3D image or videos as inputs and output the labels of objects in  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon.  ... 
doi:10.1109/ita.2016.7888183 dblp:conf/ita/ChellappaCRSKPC16 fatcat:2lbiwfm4rzfj3nlfvmo2yqn2hu

Balanced Alignment for Face Recognition: A Joint Learning Approach [article]

Huawei Wei, Peng Lu, Yichen Wei
2020 arXiv   pre-print
To strike the balance, our second contribution is a novel joint learning approach where alignment learning is controllable with respect to its strength and driven by recognition.  ...  Face alignment is crucial for face recognition and has been widely adopted. However, current practice is too simple and under-explored.  ...  Compared with these face alignment methods in preprocessing, our proposed alignment scheme does not rely on manual designed template and can be trained end-to-end. In Learning Jaderberg et al.  ... 
arXiv:2003.10168v1 fatcat:uo3pxm7klza77koyz74vxumthu

Rethinking Feature Discrimination and Polymerization for Large-scale Recognition [article]

Yu Liu and Hongyang Li and Xiaogang Wang
2017 arXiv   pre-print
COCO is bundled with discriminative training and learned end-to-end with stable convergence.  ...  To this end, we proposed the congenerous cosine (COCO) algorithm to simultaneously optimize the cosine similarity among data.  ...  COCO can be learned in a neat way with stable end-to-end training.  ... 
arXiv:1710.00870v2 fatcat:qsu3waacuzbdjolani35jarrjm

Towards Large-Pose Face Frontalization in the Wild [article]

Xi Yin, Xiang Yu, Kihyuk Sohn, Xiaoming Liu, Manmohan Chandraker
2017 arXiv   pre-print
Despite recent advances in face recognition using deep learning, severe accuracy drops are observed for large pose variations in unconstrained environments.  ...  Incorporating 3DMM into the GAN structure provides shape and appearance priors for fast convergence with less training data, while also supporting end-to-end training.  ...  ∃y h f − h 2 2 , y. (9) To summarize the framework, the reconstruction module R provides guidance to the frontalization through (2), the discriminator does so through (6) and the recognition engine through  ... 
arXiv:1704.06244v3 fatcat:dkzajxxy2fhmnfwxjodvywy4va

Investigating Nuisance Factors in Face Recognition with DCNN Representation

Claudio Ferrari, Giuseppe Lisanti, Stefano Berretti, Alberto Del Bimbo
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
Deep learning based approaches proved to be dramatically effective to address many computer vision applications, including "face recognition in the wild".  ...  In this work, we evaluate the effect of different bounding box dimensions, alignment, positioning and data source on face recognition using DCNNs, and present a thorough evaluation on two well known, public  ...  Government is authorized to reproduce and distribute reprints for Governmental purpose notwithstanding any copyright annotation thereon.  ... 
doi:10.1109/cvprw.2017.86 dblp:conf/cvpr/FerrariLBB17 fatcat:5y6gvq5vkjegnjvroyua5q6ize

Integrated Face Analytics Networks through Cross-Dataset Hybrid Training [article]

Jianshu Li, Shengtao Xiao, Fang Zhao, Jian Zhao, Jianan Li, Jiashi Feng, Shuicheng Yan, Terence Sim
2017 arXiv   pre-print
In this paper we propose an integrated Face Analytics Network (iFAN), which is able to perform multiple tasks jointly for face analytics with a novel carefully designed network architecture to fully facilitate  ...  It consists of a number of tasks, such as facial emotion recognition and face parsing, and most existing approaches generally treat these tasks independently, which limits their deployment in real scenarios  ...  , we rely on the task interaction to perform alignment-free emotion recognition.  ... 
arXiv:1711.06055v1 fatcat:f5recx7rwnecjmilttiruptbk4

Expression Recognition in the Wild Using Sequence Modeling [article]

Sowmya Rasipuram, Junaid Hamid Bhat, Anutosh Maitra
2020 arXiv   pre-print
Expression recognition in the wild is a very interesting problem and is challenging as it involves detailed feature extraction and heavy computation.  ...  As we exceed upon the procedures for modelling the different aspects of behaviour, expression recognition has become a key field of research in Human Computer Interactions.  ...  An ene-to-end model is trained to perform per-frame emotion recognition using Gated Recurrent Unit. We observed significant improvement over baseline using multi-modal approach on validation data.  ... 
arXiv:2003.00170v1 fatcat:ufab3y66irf7jmmoivhbr7he64

A CNN Model for Head Pose Recognition using Wholes and Regions

Ardhendu Behera, Andrew G Gidney, Zachary Wharton, Daniel Robinson, Keiron Quinn
2019 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019)  
In this paper, we present a novel end-to-end deep network, which is inspired by these ideas and explores regions within an image to capture topological changes due to changes in viewpoint.  ...  Currently, head pose is often computed by localizing landmarks on a targeted face and solving 2D to 3D correspondence problem with a mean head model.  ...  They use a simple CNN to regress 3D head poses, focusing on facial alignment using the predicted head pose. The facial alignment pipeline is targeted towards improving the face recognition accuracy.  ... 
doi:10.1109/fg.2019.8756536 dblp:conf/fgr/BeheraGWRQ19 fatcat:wxvp27qczndc5ad2edjj5dd7jm

Landmark Detection in Low Resolution Faces with Semi-Supervised Learning [article]

Amit Kumar, Rama Chellappa
2019 arXiv   pre-print
This deters the performance of algorithms relying on quality landmarks, for example, face recognition.  ...  recognition in low resolution images.  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon.  ... 
arXiv:1907.13255v1 fatcat:wiol6ma3rfgz3hf7lpoqg2tubi

Pose-Aware Face Recognition in the Wild

Iacopo Masi, Stephen Rawls, Gerard Medioni, Prem Natarajan
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We propose a method to push the frontiers of unconstrained face recognition in the wild, focusing on the problem of extreme pose variations.  ...  As opposed to current techniques which either expect a single model to learn pose invariance through massive amounts of training data, or which normalize images to a single frontal pose, our method explicitly  ...  We would like to thank all the other authors of [1] for their effort on the project. Moreover, thank A. D. Bagdanov, L.  ... 
doi:10.1109/cvpr.2016.523 dblp:conf/cvpr/MasiRMN16 fatcat:76jbaoszkfagtoqgafuujut7ri

GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstruction [article]

Baris Gecer, Stylianos Ploumpis, Irene Kotsia, Stefanos Zafeiriou
2019 arXiv   pre-print
We optimize the parameters with the supervision of pretrained deep identity features through our end-to-end differentiable framework.  ...  The texture features either correspond to components of a linear texture space or are learned by auto-encoders directly from in-the-wild images.  ...  We optimize the parameters with the supervision of pretrained deep identity features through our end-to-end differentiable framework.  ... 
arXiv:1902.05978v2 fatcat:uw2gxh3onbdzzaogl44ivjq3ja

Integrated Face Analytics Networks through Cross-Dataset Hybrid Training

Jianshu Li, Shengtao Xiao, Fang Zhao, Jian Zhao, Jianan Li, Jiashi Feng, Shuicheng Yan, Terence Sim
2017 Proceedings of the 2017 ACM on Multimedia Conference - MM '17  
In this paper we propose an integrated Face Analytics Network (iFAN), which is able to perform multiple tasks jointly for face analytics with a novel carefully designed network architecture to fully facilitate  ...  It consists of a number of tasks, such as facial emotion recognition and face parsing, and most existing approaches generally treat these tasks independently, which limits their deployment in real scenarios  ...  , we rely on the task interaction to perform alignment-free emotion recognition.  ... 
doi:10.1145/3123266.3123438 dblp:conf/mm/LiXZZLFYS17 fatcat:l6ykbudao5df5ndijg6zk7ihj4
« Previous Showing results 1 — 15 out of 132,787 results