2,565 Hits in 7.2 sec

A collaborative representation based projections method for feature extraction

Wankou Yang, Zhenyu Wang, Changyin Sun
2015 Pattern Recognition  
In graph embedding based methods, we usually need to manually choose the nearest neighbors and then compute the edge weights using the nearest neighbors via L2 norm (e.g. LLE).  ...  CRP is much faster than SPP since CRP calculates the objective function with L2 norm while SPP calculate the objective function with L1 norm.  ...  In CRP, for classification, we consider two quantities: local information and global information in the modeling process. The local quantity is characterized via the L2 norm graph.  ... 
doi:10.1016/j.patcog.2014.07.009 fatcat:uyl4kfqhzfb7rbh7rqcerklio4

Learning efficient structured dictionary for image classification [article]

Zi-Qi Li, Jun Sun, Xiao-Jun Wu, He-Feng Yin
2020 arXiv   pre-print
In this paper, we present an efficient structured dictionary learning (ESDL) method which takes both the diversity and label information of training samples into account.  ...  Moreover, in contrast with conventional DL approaches which impose computationally expensive L1-norm constraint on the coefficient matrix, ESDL employs L2-norm regularization term.  ...  Fig 10 10 Recognition accuracy of ESDL versus parameters α and γ on the LFW database. 20 We proposed an efficient structured dictionary learning (ESDL) method in which both the diversity and label information  ... 
arXiv:2002.03271v2 fatcat:knxm7mcxebgurihjmxigwl333m

Simultaneous Object Recognition and Localization in Image Collections

Shao-Chuan Wang, Yu-Chiang Frank Wang
2010 2010 7th IEEE International Conference on Advanced Video and Signal Based Surveillance  
via two different types of support vector machines: the standard L2regularized L1-loss SVM, and the one with L1 regularization and L2 loss.  ...  We show that the use of our visual attention maps improves the recognition performance, while the one selected by L1-regularized L2-loss SVMs exhibits the best recognition and localization results.  ...  Acknowledgments This work is supported by NSC98-2218-E-001-004 and NSC99-2631-H-001-018.  ... 
doi:10.1109/avss.2010.47 dblp:conf/avss/WangW10 fatcat:7tllmbyojncsfiss4ux4gezhta

A Novel Approach for Stable Selection of Informative Redundant Features from High Dimensional fMRI Data [article]

Yilun Wang, Zhiqiang Li, Yifeng Wang, Xiaona Wang, Junjie Zheng, Xujuan Duan, Huafu Chen
2016 arXiv   pre-print
In order to improve the stability, generalization and interpretations of the discovered potential biomarker and enhance the robustness of the resultant classifier, the redundant but informative features  ...  However, traditional multivariate methods is likely to obtain unstable and unreliable results in case of an extremely high dimensional feature space and very limited training samples, where the features  ...  our method together with L2-SVM are the only two methods which can find out the accurately discriminative regions.  ... 
arXiv:1506.08301v2 fatcat:l5rhio2wbjdj7esmfmoebbcks4

l2, 1 Regularized correntropy for robust feature selection

Ran He, Tieniu Tan, Liang Wang, Wei-Shi Zheng
2012 2012 IEEE Conference on Computer Vision and Pattern Recognition  
Extensive experiments show that our method can select robust and sparse features, and outperforms several state-of-the-art subspace methods on largescale and open face recognition datasets.  ...  In terms of face recognition, we apply the proposed method to obtain an appearance-based model, called Sparse-Fisherfaces.  ...  Since discriminative LPP is based on a local structure which depends on label information, it is sensitive to mislabeling noise.  ... 
doi:10.1109/cvpr.2012.6247966 dblp:conf/cvpr/HeTWZ12 fatcat:b2cafx7n7ngndnie4g4drfg7ue

Structure Preserving Low-Rank Representation for Semi-supervised Face Recognition [chapter]

Yong Peng, Suhang Wang, Shen Wang, Bao-Liang Lu
2013 Lecture Notes in Computer Science  
Constructing an informative and discriminative graph plays an important role in the graph based semi-supervised learning methods.  ...  Among these graph construction methods, low-rank representation based graph, which calculates the edge weights of both labeled and unlabeled samples as the low-rank representation (LRR) coefficients, has  ...  As a result, the proposed model derives a informative graph and shows the best performance in the comparison with state-of-the-art methods for semi-supervised face recognition.  ... 
doi:10.1007/978-3-642-42042-9_19 fatcat:zfp3k66a2rge7dpcri5yyzb6my

Dimension reduction using collaborative representation reconstruction based projections

Juliang Hua, Huan Wang, Mingwu Ren, Heyan Huang
2016 Neurocomputing  
KPCA is nonlinear version of PCA via kernel tricks and could handle implicitly deal with the nonlinear data. KFLDA is the nonlinear version of FLDA via kernel tricks. J. Yang  ...  So the proposed method is called CRRP. The experimental results on AR, Yale B and CMU PIE face databases demonstrate that CRRP is an effective dimension reduction method.  ...  CRC is a regularized least square regression with L2-norm constraint. C. Ren et al. [41] proposed a L21-norm-based regression 3 based classification method. F. Shen et al.  ... 
doi:10.1016/j.neucom.2016.01.060 fatcat:umdatgweyndchdb7nzwwerp6qa

Feature Transfer Learning for Deep Face Recognition with Under-Represented Data [article]

Xi Yin, Xiang Yu, Kihyuk Sohn, Xiaoming Liu, Manmohan Chandraker
2019 arXiv   pre-print
Despite the large volume of face recognition datasets, there is a significant portion of subjects, of which the samples are insufficient and thus under-represented.  ...  Advantageous results on LFW, IJB-A and MS-Celeb-1M demonstrate the effectiveness of our feature transfer and training strategy, compared to both general baselines and state-of-the-art methods.  ...  Train↓ MethodRegular Long-tail Regular Long-tail Regular Long-tail Test → MS1M: FC MS1M: NN Weight Norm 10K0K sfmx+m-L2 92.03 - 90.21 84.64 0.427 - 10K10K sfmx+m-L2 90.76 0.15 89.48 84.10  ... 
arXiv:1803.09014v2 fatcat:jnyxg7qxwralvprypm2txk6xdi

Sparse Network-Based Models for Patient Classification Using fMRI

Maria J. Rosa, Liana Portugal, John Shawe-Taylor, Janaina Mourao-Miranda
2013 2013 International Workshop on Pattern Recognition in Neuroimaging  
Here, we build a novel sparse network-based discriminative modeling framework, based on Gaussian graphical models and L1-norm regularized linear Support Vector Machines (SVM).  ...  gender discrimination and emotional task, respectively, during the visualization of emotional valent faces.  ...  Introduction Recent research using pattern recognition methods applied to whole-brain neuroimaging data, such as structural/functional Magnetic Resonance Imaging (s/fMRI) data, has proved successful at  ... 
doi:10.1109/prni.2013.26 dblp:conf/prni/RosaPSM13 fatcat:sqyrhj7jubgm5fmfafxlidnjla

Sparse network-based models for patient classification using fMRI

Maria J. Rosa, Liana Portugal, Tim Hahn, Andreas J. Fallgatter, Marta I. Garrido, John Shawe-Taylor, Janaina Mourao-Miranda
2015 NeuroImage  
Here, we build a novel sparse network-based discriminative modeling framework, based on Gaussian graphical models and L1-norm regularized linear Support Vector Machines (SVM).  ...  gender discrimination and emotional task, respectively, during the visualization of emotional valent faces.  ...  Introduction Recent research using pattern recognition methods applied to whole-brain neuroimaging data, such as structural/functional Magnetic Resonance Imaging (s/fMRI) data, has proved successful at  ... 
doi:10.1016/j.neuroimage.2014.11.021 pmid:25463459 pmcid:PMC4275574 fatcat:w3wdj2ll7jamlcy5hfyjjbag34

L2-constrained Softmax Loss for Discriminative Face Verification [article]

Rajeev Ranjan, Carlos D. Castillo, Rama Chellappa
2017 arXiv   pre-print
A typical pipeline for face verification includes training a deep network for subject classification with softmax loss, using the penultimate layer output as the feature descriptor, and generating a cosine  ...  Additionally, we achieve state-of-the-art performance on LFW dataset with an accuracy of 99.78%, and competing performance on YTF dataset with accuracy of 96.08%.  ...  (a) Face Verification Performance on IJB-A dataset. The templates are divided into 3 sets based on their L2-norm. '1' denotes the set with low L2-norm while '3' represents high L2-norm.  ... 
arXiv:1703.09507v3 fatcat:osuh5c62c5g7tcfd5ijhbvrilu

Random Faces Guided Sparse Many-to-One Encoder for Pose-Invariant Face Recognition

Yizhe Zhang, Ming Shao, Edward K. Wong, Yun Fu
2013 2013 IEEE International Conference on Computer Vision  
Finally, we conduct face identification on CMU Multi-PIE, and verification on Labeled Faces in the Wild (LFW) databases, where identification rank-1 accuracy and face verification accuracy with ROC curve  ...  One of the most challenging task in face recognition is to identify people with varied poses. Namely, the test faces have significantly different poses compared with the registered faces.  ...  This research is supported in part by the NSF CNS award 1314484, Office of Naval Research award N00014-12-1-0125 and N00014-12-1-1028, Air Force Office of Scientific Research award FA9550-12-1-0201, and  ... 
doi:10.1109/iccv.2013.300 dblp:conf/iccv/ZhangSWF13 fatcat:qyt77qgmabhvvjntaa3g2o3hnm

Robust human face recognition based on locality preserving sparse over complete block approximation

Dimche Kostadinov, Svyatoslav Voloshynovskiy, Sohrab Ferdowsi, Adnan M. Alattar, Nasir D. Memon, Chad D. Heitzenrater
2014 Media Watermarking, Security, and Forensics 2014  
Compressive Sensing (CS) has become one of the standard methods in face recognition due to the success of the family of Sparse Representation based Classification (SRC) algorithms.  ...  We compare two image representations using a pixel-wise approximation and an overcomplete block-wise approximation with two types of sparsity priors.  ...  The authors are thankful to Maurits Diephuis for the discussions on the links to the deep learning encoding algorithms and his comments in the early versions of the paper.  ... 
doi:10.1117/12.2042506 dblp:conf/mediaforensics/KostadinovVF14 fatcat:zx6d2olkozgxvo7mhosn324mkm

Locality Preserving and Label-Aware Constraint-Based Hybrid Dictionary Learning for Image Classification

Jianqiang Song, Lin Wang, Zuozhi Liu, Muhua Liu, Mingchuan Zhang, Qingtao Wu
2021 Applied Sciences  
Moreover, all the introduced constraints in the proposed LPLC-HDL method are based on the l2-norm regularization, which can be solved efficiently via employing an alternative optimization strategy.  ...  In this paper, we proposed a locality preserving and label-aware constraint-based hybrid dictionary learning (LPLC-HDL) method, and apply it in image classification effectively.  ...  make the coding coefficients V p have a block-diagonal structure with strong discriminative information.  ... 
doi:10.3390/app11167701 fatcat:wzcundxrpze4xj7yyawum6auda

Graph Regularized Nonnegative Matrix Factorization with Sparse Coding

Chuang Lin, Meng Pang
2015 Mathematical Problems in Engineering  
Experimental results demonstrate encouraging results of GRNMF_SC on image recognition and clustering when comparing with the other state-of-the-art NMF methods.  ...  In this paper, we propose a sparseness constraint NMF method, named graph regularized matrix factorization with sparse coding (GRNMF_SC).  ...  Acknowledgments Chuang Lin acknowledges the financial support of Natural Science Foundation of China (no. 61272371) and Fundamental Research Funds for the Central Universities, Dalian University of Technology  ... 
doi:10.1155/2015/239589 fatcat:mvc2iqvrt5a5jd3m2l2qrgq7ie
« Previous Showing results 1 — 15 out of 2,565 results