Filters








45,823 Hits in 5.4 sec

Manifold Modeling with Learned Distance in Random Projection Space for Face Recognition

Grigorios Tsagkatakis, Andreas Savakis
2010 2010 20th International Conference on Pattern Recognition  
Manifold learning describes the process by which this low dimensional embedding can be generated.  ...  We demonstrate that this approach is effective for multi view face recognition.  ...  Once the manifold embedding is learned, classification of new data points is achieved by measuring the learned distance between the embedded training samples and the new data point.  ... 
doi:10.1109/icpr.2010.165 dblp:conf/icpr/TsagkatakisS10 fatcat:zlixruu6ijbl7bmd2lzqi2ijzu

Spherical and Hyperbolic Embeddings of Data

Richard C. Wilson, Edwin R. Hancock, Elzbieta Pekalska, Robert P. W. Duin
2014 IEEE Transactions on Pattern Analysis and Machine Intelligence  
In each case the embedding maintains the local structure of the data while placing the points in a metric space.  ...  Many computer vision and pattern recognition problems may be posed as the analysis of a set of dissimilarities between objects.  ...  The embedding space is non-metric and the squared-distance between pairs of points in the space can be negative.  ... 
doi:10.1109/tpami.2014.2316836 pmid:26353065 fatcat:kb4gqcud3fcazmduukttjllmpm

On Learning Density Aware Embeddings

Soumyadeep Ghosh, Richa Singh, Mayank Vatsa
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
The proposed method, termed as Density Aware Metric Learning, enforces the model to learn embeddings that are pulled towards the most dense region of the clusters for each class.  ...  Deep metric learning algorithms have been utilized to learn discriminative and generalizable models which are effective for classifying unseen classes.  ...  learning method for effective training especially with noisy data. • Detailed analysis and comparison with other popular deep metric learning methods on four challenging databases pertaining to face and  ... 
doi:10.1109/cvpr.2019.00502 dblp:conf/cvpr/Ghosh0V19 fatcat:azpqszdysbdtrkvpaxmq75qbhq

Viewpoint Manifolds for Action Recognition

Richard Souvenir, Kyle Parrigan
2009 EURASIP Journal on Image and Video Processing  
In real-world settings, the viewpoint of the camera cannot always be fixed relative to the subject, so view-invariant action recognition methods are needed.  ...  In this paper, we present a framework for learning a compact representation of primitive actions (e.g., walk, punch, kick, sit) that can be used for video obtained from a single camera for simultaneous  ...  Isomap embeds points in a low-dimensional Euclidean space by preserving the geodesic pair-wise distances of the points in original space.  ... 
doi:10.1155/2009/738702 fatcat:h4anyadp3be6rojbcbur4oy4pa

Conformal Embedding Analysis with Local Graph Modeling on the Unit Hypersphere

Yun Fu, Ming Liu, Thomas S. Huang
2007 2007 IEEE Conference on Computer Vision and Pattern Recognition  
We present the Conformal Embedding Analysis (CEA) for feature extraction and dimensionality reduction.  ...  The subspace learned by CEA is graylevel variation tolerable since the cosine-angle metric and the normalization processing enhance the robustness of the conformal feature extraction.  ...  Since CEA is a linear subspace learning method 2 , we compare it with PCA [8] , LDA [9] , LPP [5] , and LSDA [27] , four most popular linear methods in face recognition.  ... 
doi:10.1109/cvpr.2007.383410 dblp:conf/cvpr/FuLH07 fatcat:tgeqabld3nh5hn6bgrjuh57gg4

Spam image discrimination using support vector machine based on higher-order local autocorrelation feature extraction

Hongrong Cheng, Zhiguang Qin, Qiao Liu, Mingcheng Wan
2008 2008 IEEE Conference on Cybernetics and Intelligent Systems  
Experimental results for the public personal dataset show that the proposed method can separate spam images from non-spam images with minimum recognition rates of 98%.  ...  This method extracts edge features of a binarized image by using higher-order local autocorrelation(HLAC), and then input those features to support vector machine (SVM) for classification.  ...  Fig. 5 . 5 Classification precision for the four kernels Fig. 6 . 6 Classification recall for the four kernels TABLE I I RECOGNITION RATES RELATED METRICS Real spam(t=1) Real non-spam(t=-1) Predicted  ... 
doi:10.1109/iccis.2008.4670821 fatcat:ex3n3glpo5ezvjjbrd5m4fe5j4

Special issue on Graph-Based Representations in Computer Vision

Edwin R. Hancock, Andrea Torsello, Francisco Escolano, Luc Brun
2011 Computer Vision and Image Understanding  
''Graph Attribute Embedding via Riemannian Submersion Learning'' by Haifeng Zhao et al. studies the problem of embedding a set of relational structures into a Riemmanian metric space.  ...  into a vectorial space and recursively recovering a graph from a point in such space.  ... 
doi:10.1016/j.cviu.2011.04.003 fatcat:ttotuzgpgzbxnpkxkbjlxspvyq

Relaxational metric adaptation and its application to semi-supervised clustering and content-based image retrieval

Hong Chang, Dit-Yan Yeung, William K. Cheung
2006 Pattern Recognition  
Recently, some novel methods (e.g., the parametric method proposed by Xing et al.) for learning global metrics based on pairwise side information have been shown to demonstrate promising results.  ...  In this paper, we propose a nonparametric method, called relaxational metric adaptation (RMA), for the same metric adaptation problem.  ...  Acknowledgments The research described in this paper has been supported by two grants, CA03/04.EG01 (which is part of HKBU2/03/C) and HKUST6174/04E, from the Research Grants Council of the Hong Kong Special  ... 
doi:10.1016/j.patcog.2006.04.006 fatcat:4pd5rvnqlbdvxltyu2f4u3jdu4

On Learning Density Aware Embeddings [article]

Soumyadeep Ghosh, Richa Singh, Mayank Vatsa
2019 arXiv   pre-print
The proposed method, termed as Density Aware Metric Learning, enforces the model to learn embeddings that are pulled towards the most dense region of the clusters for each class.  ...  Deep metric learning algorithms have been utilized to learn discriminative and generalizable models which are effective for classifying unseen classes.  ...  learning method for effective training especially with noisy data. • Detailed analysis and comparison with other popular deep metric learning methods on four challenging databases pertaining to face and  ... 
arXiv:1904.03911v1 fatcat:izztuyi74beitnq6ipawawcrfq

Rethinking Task and Metrics of Instance Segmentation on 3D Point Clouds [article]

Kosuke Arase, Yusuke Mukuta, Tatsuya Harada
2019 arXiv   pre-print
Our method learns a mapping from input point clouds to an embedding space, where the embeddings form clusters for each instance and distinguish instances using these clusters during testing.  ...  To address these problems, we propose a new method with space complexity O(Np) such that large regions can be consumed, as well as novel metrics for tasks that are independent of the categories or size  ...  We proposed a new method for instance segmentation on 3D point clouds. Our memory efficient loss function learns mapping to the embedding space, where the embeddings form clusters for each object.  ... 
arXiv:1909.12655v1 fatcat:37luof44lngzlihla3p6scilye

Comparison of Multidimensional Data Access Methods for Feature-Based Image Retrieval

Serdar Arslan, Ahmet Sacan, Esra Acar, I. Hakki Toroslu, Adnan Yazici
2010 2010 20th International Conference on Pattern Recognition  
The time and accuracy trade-offs for each of these methods are demonstrated on a large Corel image database.  ...  We further show that using multidimensional scaling can achieve comparable accuracy, while speeding-up the query times significantly by allowing the use of spatial access methods.  ...  , which is an index structure on metric spaces [3] , and a Landmark Multidimensional Scaling (LMDS) method which uses FastMap method for landmark selection [4] .  ... 
doi:10.1109/icpr.2010.797 dblp:conf/icpr/ArslanSATY10 fatcat:nn6kveedvbdmldcnv32kupvque

Penalized K-Nearest-Neighbor-Graph Based Metrics for Clustering [article]

Ariel E. Baya, Pablo M. Granitto
2010 arXiv   pre-print
We use three artificial datasets in four different embedding situations to evaluate the behavior of the new metric, including a comparison among different clustering methods.  ...  weight for connecting the sub-graphs produced by the first step.  ...  Verdes for useful comments on this and on a previous manuscript, and Monica Larese and Lucas Uzal for a carefull reading of this manuscript.  ... 
arXiv:1006.2734v1 fatcat:ruihvkwvtncdhhbidvfjwyqvai

Non-Euclidean properties of spike train metric spaces

Dmitriy Aronov, Jonathan D. Victor
2004 Physical Review E  
Our results demonstrate that metric spaces enrich the study of neural activity patterns, since accounting for perceptual spaces requires a non-Euclidean geometry.  ...  In neuroscience, traditional methods for sequence comparisons rely on techniques appropriate for multivariate data, which typically assume that the space of sequences is intrinsically Euclidean.  ...  ACKNOWLEDGMENTS We thank Tom Schneider and Marcelo Magnasco for comments on the manuscript. This work was supported in part by NEI EY9314.  ... 
doi:10.1103/physreve.69.061905 pmid:15244615 pmcid:PMC2911631 fatcat:4j5xwp6r55gwzook7h5o2ecrs4

Radar-PointGNN: Graph Based Object Recognition for Unstructured Radar Point-cloud Data

Peter Svenningsson, Francesco Fioranelli, Alexander Yarovoy
2021 2021 IEEE Radar Conference (RadarConf21)  
An object recognition model is here presented which imposes a graph structure on the radar point-cloud by connecting spatially proximal points and extracts local patterns by performing convolutional operations  ...  While object recognition systems in the LiDAR and camera modalities are reaching maturity, recognition models on sparse radar point measurements have remained an open research challenge.  ...  I also thank nuTonomy for enabling a large body of research by making the nuScenes dataset publicly available.  ... 
doi:10.1109/radarconf2147009.2021.9455172 fatcat:6thxt5wmezfqvcazqoq7i652zq

Discriminant Neighborhood Structure Embedding Using Trace Ratio Criterion for Image Recognition

Jing Wang, Fang Chen, Quanxue Gao
2015 Journal of Computer and Communications  
Dimensionality reduction is very important in pattern recognition, machine learning, and image recognition.  ...  the sum of the weighted distances between nearby points from different classes.  ...  points in the low dimensional space.  ... 
doi:10.4236/jcc.2015.311011 fatcat:u6g2qtslbbehbgmqnpbatgypy4
« Previous Showing results 1 — 15 out of 45,823 results