Filters








43,746 Hits in 5.4 sec

Learning structurally discriminant features in 3D faces

Sreenivas R. Sukumar, Hamparsum Bozdogan, David L. Page, Andreas F. Koschan, Mongi A. Abidi
2008 2008 15th IEEE International Conference on Image Processing  
In this paper, we derive a data mining framework to analyze 3D features on human faces.  ...  Index Terms-3D face recognition, feature learning, dimensionality reduction, informative-discrimant face features.  ...  The application of the framework to learn structurally diverse, discriminant and informative 3D facial features extracted from a database of 300 faces consisting of people from different gender and ethnic  ... 
doi:10.1109/icip.2008.4712154 dblp:conf/icip/SukumarBPKA08 fatcat:jraejddabrcpvnt424lq2vfgui

CP-GAN: A Cross-Pose Profile Face Frontalization Boosting Pose-Invariant Face Recognition

Jinjin Liu, Qingbao Li, Ming Liu, Tongxin Wei
2020 IEEE Access  
In PIM, a complex discriminative learning sub-net is adopted for face verification and recognition.  ...  Although synthesized faces have a similar color tone with images from CASIA 3D FACE, CP-GAN can faithfully frontalize profile faces in IJB-A with both fine facial details and proper global structure.  ... 
doi:10.1109/access.2020.3033675 fatcat:enjgquxsfng7pmrpyheosh4pby

2D+3D Facial Expression Recognition via Discriminative Dynamic Range Enhancement and Multi-Scale Learning [article]

Yang Jiao, Yi Niu, Trac D. Tran, Guangming Shi
2020 arXiv   pre-print
In 2D+3D facial expression recognition (FER), existing methods generate multi-view geometry maps to enhance the depth feature representation.  ...  Finally, we also design an efficient Facial Attention structure to automatically locate subtle discriminative facial parts for multi-scale learning, and train it with a proposed loss function ℒ_FA without  ...  Once accomplished, multi-scale feature learning via Facial Attention structure is employed in the final recognition stage.  ... 
arXiv:2011.08333v1 fatcat:wapprzdrobhlho3j3ygsckqmsq

Coupled Discriminative Feature Learning for Heterogeneous Face Recognition

Yi Jin, Jiwen Lu, Qiuqi Ruan
2015 IEEE Transactions on Information Forensics and Security  
The goal of CDFL is to learn an optimal filter α, which makes the feature of filtered images more discriminative. • Our principle of feature learning in CDFL: The principle of feature learning in CDFL.  ...  The basic idea of feature learning to seek optimal image filters to obtain discriminative face representation.  ... 
doi:10.1109/tifs.2015.2390414 fatcat:aajcbsougjfxdd6jjyuhevpn2e

Deep Geometric Texture Synthesis [article]

Amir Hertz, Rana Hanocka, Raja Giryes, Daniel Cohen-Or
2020 arXiv   pre-print
It learns deep features on the faces of the input triangulation, which is used to subdivide and generate offsets across multiple scales, without parameterization of the reference or target mesh.  ...  It learns geometric texture statistics from local neighborhoods (i.e., local triangular patches) of a single reference 3D model.  ...  In the same spirit, in this work, we learn the distribution of local patches, but of 3D triangular meshes, which, unlike images, have an irregular structure. Deep generative models in 3D.  ... 
arXiv:2007.00074v1 fatcat:zz5faywvr5fo7no5cvvwfrdcka

3D-Aided Deep Pose-Invariant Face Recognition

Jian Zhao, Lin Xiong, Yu Cheng, Yi Cheng, Jianshu Li, Li Zhou, Yan Xu, Jayashree Karlekar, Sugiri Pranata, Shengmei Shen, Junliang Xing, Shuicheng Yan (+1 others)
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
To mitigate this gap, we propose a 3D-Aided Deep Pose-Invariant Face Recognition Model (3D-PIM), which automatically recovers realistic frontal faces from arbitrary poses through a 3D face model in a novel  ...  Specifically, 3D-PIM incorporates a simulator with the aid of a 3D Morphable Model (3D MM) to obtain shape and appearance prior for accelerating face normalization learning, requiring less training data  ...  If one visualizes the learned deep features in the high-dimensional space, these learned deep features form several compact clusters, and each cluster may be far away from others.  ... 
doi:10.24963/ijcai.2018/165 dblp:conf/ijcai/ZhaoXCCLZXKPSXY18 fatcat:2bfsdg742veahofrzjsped5zje

Cross-Modality 2D-3D Face Recognition via Multiview Smooth Discriminant Analysis Based on ELM

Yi Jin, Jiuwen Cao, Qiuqi Ruan, Xueqiao Wang
2014 Journal of Electrical and Computer Engineering  
In this paper, we propose a new approach for cross-modality 2D-3D face recognition (FR), which is called Multiview Smooth Discriminant Analysis (MSDA) based on Extreme Learning Machines (ELM).  ...  In recent years, 3D face recognition has attracted increasing attention from worldwide researchers.  ...  feature space for the cross-modality 2D-3D face features.  ... 
doi:10.1155/2014/584241 fatcat:gqjvpddltnda7ngryopqfyuwti

Texturify: Generating Textures on 3D Shape Surfaces [article]

Yawar Siddiqui, Justus Thies, Fangchang Ma, Qi Shan, Matthias Nießner, Angela Dai
2022 arXiv   pre-print
In particular, our method does not require any 3D color supervision or correspondence between shape geometry and images to learn the texturing of 3D objects.  ...  We thus propose Texturify, a GAN-based method that leverages a 3D shape dataset of an object class and learns to reproduce the distribution of appearances observed in real images by generating high-quality  ...  on Static and Dynamic 3D Data Practical.  ... 
arXiv:2204.02411v1 fatcat:m3z2xhadkzhy7meng3ef5i75qa

View Independent Generative Adversarial Network for Novel View Synthesis

Xiaogang Xu, Yingcong Chen, Jiaya Jia
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
To this end, an encoder is designed to extract view-independent feature that characterizes intrinsic properties of the input image, which includes 3D structure, color, texture etc.  ...  Synthesizing novel views from a 2D image requires to infer 3D structure and project it back to 2D from a new viewpoint.  ...  VI-GAN is trained with weakly supervised 2D data, while learned features are beneficial to 3D-related learning tasks. Figure 1 . 1 Overall structure of VI-GAN.  ... 
doi:10.1109/iccv.2019.00788 dblp:conf/iccv/XuCJ19 fatcat:xh5c6d4mh5dnzaajg2dpddlaae

2D-3D Heterogeneous Face Recognition Based on Deep Coupled Spectral Regression

Yangtao Zheng, Di Huang, Weixin Li, Shupeng Wang, Yunhong Wang
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
As one of the major branches in Face Recognition (FR), 2D-3D Heterogeneous FR (HFR), where face comparison is achieved across the texture and shape modalities, has become more important.  ...  Specifically, from 2D texture and 3D depth face maps, DCSR extracts more powerful features by a deep network with the cross-modality triplet loss, which show much better uniqueness and robustness than  ...  For joint optimization in feature extraction and mapping learning, we build a simple yet effective layer structure, called the couple layer. The overall structure is shown in Figure 2 .  ... 
doi:10.1109/cvprw.2019.00037 dblp:conf/cvpr/Zheng0LWW19 fatcat:i7k6cg3zhrehlmzfiiggbp7oce

Adversarial Learning of Structure-Aware Fully Convolutional Networks for Landmark Localization [article]

Yu Chen, Chunhua Shen, Hao Chen, Xiu-Shen Wei, Lingqiao Liu, Jian Yang
2019 arXiv   pre-print
If the pose generator G generates results that the discriminator fails to distinguish from real ones, the network successfully learns the priors.  ...  The proposed approach significantly outperforms the state-of-the-art methods and almost always generates plausible pose predictions, demonstrating the usefulness of implicit learning of structures using  ...  This work was in part supported by an ARC Future Fellowship to C. Shen and an ARC DECRA Fellowship to L. Liu; and ARC Grant CE140100016.  ... 
arXiv:1711.00253v5 fatcat:3laenudkwjbqleiszofe6paxja

Adversarial Learning of Structure-Aware Fully Convolutional Networks for Landmark Localization

Yu Chen, Chunhua Shen, Hao Chen, Xiu-Shen Wei, Lingqiao Liu, Jian Yang
2019 IEEE Transactions on Pattern Analysis and Machine Intelligence  
If the pose generator G generates results that the discriminator fails to distinguish from real ones, the network successfully learns the priors.  ...  The proposed approach significantly outperforms several state-of-the-art methods and almost always generates plausible pose predictions, demonstrating the usefulness of implicit learning of structures  ...  This work was in part supported by an ARC Future Fellowship to C. Shen; and an ARC DECRA Fellowship to L. Liu.  ... 
doi:10.1109/tpami.2019.2901875 pmid:30835211 fatcat:bwcawevazfcp3dlk5qcg4j3ube

Robust 3D Face Recognition Using Learned Visual Codebook

Cheng Zhong, Zhenan Sun, Tieniu Tan
2007 2007 IEEE Conference on Computer Vision and Pattern Recognition  
In our method, we first extract intrinsic discriminative information embedded in 3D faces using Gabor filters, then K-means clustering is adopted to learn the centers from the filter response vectors.  ...  In this paper, we propose a novel learned visual codebook (LVC) for 3D face recognition.  ...  In our method we not only extract intrinsic discriminative information embedded in 3D faces using Gabor features, but also choose K-means clustering to learn basic facial elements and construct a Learned  ... 
doi:10.1109/cvpr.2007.383279 dblp:conf/cvpr/ZhongST07 fatcat:hkdhdswr6ba4jj3awj5ojgk6zq

Dissertation in UPM (Revised in 201910) SIMING ZHENG

Siming Zheng
2020 Figshare  
3D face texture recognition (Revised in 201910).  ...  CNN structures in 3D face recognition tasks.  ...  In this section, the HOG feature is used as a means of feature extraction in the process of recognition, the purpose is to combine the discriminative 3D face feature in the recognition phase, the specific  ... 
doi:10.6084/m9.figshare.11933502.v1 fatcat:77akbwkiznfsnn3ih7p2otsghy

Deep and Shallow Covariance Feature Quantization for 3D Facial Expression Recognition [article]

Walid Hariri, Nadir Farah, Dinesh Kumar Vishwakarma
2021 arXiv   pre-print
Facial expressions recognition (FER) of 3D face scans has received a significant amount of attention in recent years.  ...  A covariance matrix learning is used as a manifold layer to reduce the deep covariance matrices size and enhance their discrimination power while preserving their manifold structure.  ...  SPD matrix learning: consists of reducing the SPD matrices size and enhance their discrimination power while preserving their manifold structure.  ... 
arXiv:2105.05708v1 fatcat:hpialfljerdafe4o2eborj27a4
« Previous Showing results 1 — 15 out of 43,746 results