Filters








404 Hits in 3.0 sec

Learning Robust Graph Regularisation for Subspace Clustering

Elyor Kodirov, Tao Xiang, Zhenyong Fu, Shaogang Gong
2016 Procedings of the British Machine Vision Conference 2016   unpublished
In this work, we identify two critical limitations of the graph regularisation term employed in existing subspace clustering models and provide solutions for both of them.  ...  Various subspace clustering methods have benefited from introducing a graph regularisation term in their objective functions [2] .  ...  In this work, we identify two critical limitations of the graph regularisation term employed in existing subspace clustering models and provide solutions for both of them.  ... 
doi:10.5244/c.30.138 fatcat:acv4n7li3ndgfjqukjfplltk4y

Robust One-Class Kernel Spectral Regression [article]

Shervin Rahimzadeh Arashloo, Josef Kittler
2019 arXiv   pre-print
For both alternative regularisation schemes, iterative algorithms are proposed which recursively update label confidences and rank training observations based on their fit with the model.  ...  In this respect, first, the effect of a Tikhonov regularisation in the Hilbert space is analysed where the one-class learning problem in presence of contaminations in the training set is posed as a sensitivity  ...  approach. • DPCP is a method for learning a linear subspace from data corrupted by outliers based on a nonconvex l 1 optimisation problem [39] .  ... 
arXiv:1902.02208v1 fatcat:xxlenwx72zeptnt4ntwicoliti

Learning a Manifold as an Atlas

Nikolaos Pitelis, Chris Russell, Lourdes Agapito
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
We report state-ofthe-art results for manifold based nearest neighbour classification on vision datasets, and show how the same techniques can be applied to the 3D reconstruction of human motion from a  ...  In this work, we return to the underlying mathematical definition of a manifold and directly characterise learning a manifold as finding an atlas, or a set of overlapping charts, that accurately describe  ...  In motion tracking, [14] performed agglomerative clustering over predefined affine subspaces to learn closed manifolds.  ... 
doi:10.1109/cvpr.2013.215 dblp:conf/cvpr/PitelisRA13 fatcat:p6nld5qjzrf6zjrjrpjv5x5u2y

Segmentation of Subspaces in Sequential Data [article]

Stephen Tierney, Yi Guo, Junbin Gao
2015 arXiv   pre-print
We propose Ordered Subspace Clustering (OSC) to segment data drawn from a sequentially ordered union of subspaces.  ...  Experiments show that our method, OSC, outperforms the state of the art methods: Spatial Subspace Clustering (SpatSC), Low-Rank Representation (LRR) and SSC.  ...  Ordered Subspace Clustering The assumption for Ordered Subspace Clustering (OSC) [19] is that the data is sequentially structured.  ... 
arXiv:1504.04090v1 fatcat:iywopvz3nrb75h35pxkvyx3pre

Robust Random Walk for Leaf Segmentation

Zhibo Chen, Jing Hu, Yang Meng, Zhang Rongguo, Shuai Zhang
2020 IET Image Processing  
To address this problem, they learn a common subspace by taking into account the illumination of local and non-local pixels.  ...  A new method -robust random walk (RW) is proposed to propagate the prior of user's specified pixels.  ...  Then, we need to learn a common subspace w where the transformed features w T g are the same for similar grey patches with different illuminations.  ... 
doi:10.1049/iet-ipr.2018.6255 fatcat:qqwhtkazd5ghjjar6n3ykcge7y

Discriminative Subspace Clustering

Vasileios Zografos, Liam Ellis, Rudolf Mester
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
We present a novel method for clustering data drawn from a union of arbitrary dimensional subspaces, called Discriminative Subspace Clustering (DiSC).  ...  DiSC solves the subspace clustering problem by using a quadratic classifier trained from unlabeled data (clustering by classification).  ...  Acknowledgements We would like to thank Klas Norderg, Reiner Lenz and Michael Felsberg for the helpful discussions.  ... 
doi:10.1109/cvpr.2013.274 dblp:conf/cvpr/ZografosEM13 fatcat:6aj5734pnraevdaethysuymkc4

From Semi-supervised to Transfer Counting of Crowds

Chen Change Loy, Shaogang Gong, Tao Xiang
2013 2013 IEEE International Conference on Computer Vision  
and actively. (2) Rather than learning from only labelled data, the abundant unlabelled data are exploited. (3) Labelled data from other scenes are employed to further alleviate the burden for data annotation  ...  Regression-based techniques have shown promising results for people counting in crowded scenes. However, most existing techniques require expensive and laborious data annotation for model training.  ...  Parameter settings: The proposed method has a few free parameters, including the number of neighbours k to build the k-NN graph, the dimensionality m of PCA subspace during the determination of normal  ... 
doi:10.1109/iccv.2013.270 dblp:conf/iccv/LoyGX13 fatcat:65fnu4irxjdlnn5lf4wqge2ic4

Linear Shape Deformation Models with Local Support Using Graph-Based Structured Matrix Factorisation

Florian Bernard, Peter Gemmar, Frank Hertel, Jorge Goncalves, Johan Thunberg
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Based on matrix factorisation with sparsity and graph-based regularisation terms, we present a method to obtain deformation factors with local support.  ...  Commonly, using Principal Components Analysis a low-dimensional subspace of the high-dimensional shape space is determined.  ...  Acknowledgements We thank Yipin Yang and colleagues for making the human body shapes dataset publicly available; Benjamin D.  ... 
doi:10.1109/cvpr.2016.607 dblp:conf/cvpr/BernardGHGT16 fatcat:evstpkr22vgtdf7g4un6diqnve

Kernel Truncated Regression Representation for Robust Subspace Clustering [article]

Liangli Zhen, Dezhong Peng, Wei Wang, Xin Yao
2020 arXiv   pre-print
Subspace clustering aims to group data points into multiple clusters of which each corresponds to one subspace.  ...  and block-diagonality; and 4) executing the graph cutting operation on the coefficient matrix by solving a graph Laplacian problem.  ...  Algorithm 1 Learning kernel truncated regression representation for robust subspace clustering Input: A given data set X ∈ R m×n , the tradeoff parameter λ, the parameter η, and the number of subspaces  ... 
arXiv:1705.05108v3 fatcat:obqzayltmbco3esxpx737syvg4

What do CNN neurons learn: Visualization Clustering [article]

Haoyue Dai
2020 arXiv   pre-print
Visualization means the method of gradient descent on image pixel, and in clustering section two algorithms are proposed to cluster respectively over image categories and network neurons.  ...  Specifically, we use two techniques - visualization and clustering - to tackle the problems above.  ...  By the clustering algorithm, the image categories are mapped to a dense graph with subspace distance between them quantitatively defined, and then a hierarchical tree (like taxonomy tree) is generated.  ... 
arXiv:2010.11725v1 fatcat:dawr33okyzgu7mhq6kloicp5vu

Kernel Truncated Regression Representation for Robust Subspace Clustering

Liangli Zhen, Dezhong Peng, Wei Wang, Xin Yao
2020 Information Sciences  
[17] used a smaller graph to approximate the full graph adaptively by learning from the raw data. These methods have shown promising performance in subspace clustering.  ...  Algorithm 1 Learning kernel truncated regression representation for robust subspace clustering Input: A given data set X ∈ R m×n , the tradeoff parameter λ, the parameter η, and the number of subspaces  ... 
doi:10.1016/j.ins.2020.03.033 fatcat:m27dju426bdvblakjczmwqznvy

Robust deformable shape reconstruction from monocular video with manifold forests

Lili Tao, Bogdan J. Matuszewski
2016 Machine Vision and Applications  
The key contributions of this work are the use of random decision forests for the shape manifold learning and robust metric for calculation of the re-projection error.  ...  The learned manifold defines constraints imposed on the reconstructed shapes.  ...  Forest model for manifold learning In the proposed method, the affinity model in manifold learning is built by applying random forest clustering.  ... 
doi:10.1007/s00138-016-0769-3 fatcat:7zg3qbrrbbccdpiarxalrz6sre

Linear Shape Deformation Models with Local Support Using Graph-based Structured Matrix Factorisation [article]

Florian Bernard, Peter Gemmar, Frank Hertel, Jorge Goncalves, Johan Thunberg
2016 arXiv   pre-print
For that, based on a well-grounded theoretical motivation, we formulate a matrix factorisation problem employing sparsity and graph-based regularisation terms.  ...  Commonly, using Principal Components Analysis a low-dimensional (affine) subspace of the high-dimensional shape space is determined.  ...  Acknowledgements We thank Yipin Yang and colleagues for making the human body shapes dataset publicly available; Benjamin D. Haeffele  ... 
arXiv:1510.08291v2 fatcat:xh7rb7b5anfnrmxv44vayq3ey4

Kernel analysis on Grassmann manifolds for action recognition

Mehrtash T. Harandi, Conrad Sanderson, Sareh Shirazi, Brian C. Lovell
2013 Pattern Recognition Letters  
Modelling video sequences by subspaces has recently shown promise for recognising human actions.  ...  To achieve efficient machinery, we propose graph-based local discriminant analysis that utilises within-class and between-class similarity graphs to characterise intra-class compactness and inter-class  ...  For certain choices of the regulariser, solving (10) reduces to identifying N parameters and not the dimension of H.  ... 
doi:10.1016/j.patrec.2013.01.008 fatcat:n42extttsngaflqdmqvzxuozte

RicciNets: Curvature-guided Pruning of High-performance Neural Networks Using Ricci Flow [article]

Samuel Glass, Simeon Spasov, Pietro Liò
2020 arXiv   pre-print
The computational graph is pruned based on a node mass probability function defined by local graph measures and weighted by hyperparameters produced by a reinforcement learning-based controller neural  ...  We use the definition of Ricci curvature to remove edges of low importance before mapping the computational graph to a neural network.  ...  A smaller ratio of output degree to input degree is therefore taken to indicate better robustness with respect to graph damage.  ... 
arXiv:2007.04216v1 fatcat:yrcxstzwqvgatfgbmoqo4zygzy
« Previous Showing results 1 — 15 out of 404 results