51,980 Hits in 5.0 sec

Unsupervised Feature Selection for Multi-View Clustering on Text-Image Web News Data

Mingjie Qian, Chengxiang Zhai
2014 Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management - CIKM '14  
We propose a new multi-view unsupervised feature selection method in which image local learning regularized orthogonal nonnegative matrix factorization is used to learn pseudo labels and simultaneously  ...  State-of-theart multi-view unsupervised feature selection methods learn pseudo class labels by spectral analysis, which is sensitive to the choice of similarity metric for each view.  ...  CONCLUSION We propose a new unsupervised feature selection methods for multi-view clustering: MVUFS where local learning regularized orthogonal nonnegative matrix factorization is performed to learn pseudo  ... 
doi:10.1145/2661829.2661993 dblp:conf/cikm/QianZ14 fatcat:zva5hdsur5hkfiajtip3r4npzm

A study of change detection from satellite images using joint intensity histogram

Yasuyo Kita
2008 Pattern Recognition (ICPR), Proceedings of the International Conference on  
In this paper, we propose a method to estimate such background intensity changes by analyzing the joint intensity histogram of compared images.  ...  Detection of appearance/disappearance of objects from satellite images is generally very difficult since background pixels also change their intensity values owing to various factors.  ...  Sato of Geologic Remote Sensing Research Group, Institute for Geoscience, AIST for data provision and discussions, and also to Dr. T. Nagami, Dr. T. Masuda, and Dr. N.  ... 
doi:10.1109/icpr.2008.4761020 dblp:conf/icpr/Kita08 fatcat:wd4nwr47izf6nc4sjkfsyhyv4e

Matrix Tri-Factorization with Manifold Regularizations for Zero-Shot Learning

Xing Xu, Fumin Shen, Yang Yang, Dongxiang Zhang, Heng Tao Shen, Jingkuan Song
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Specifically, we learn the semantic embedding projection by decomposing the visual feature matrix under the guidance of semantic embedding and class label matrices.  ...  ., attributes) shared between seen and unseen classes. In this paper, we propose a novel projection framework based on matrix tri-factorization with manifold regularizations.  ...  This work was supported in part by the National Natural Science Foundation of China under Project 61602089, Project 61502081, Project 61572108, Project 61632007, and the Fundamental Research Funds for  ... 
doi:10.1109/cvpr.2017.217 dblp:conf/cvpr/XuS0ZSS17 fatcat:6msmdsf7fjcjxikwl5xdhyz5ee

JECL: Joint Embedding and Cluster Learning for Image-Text Pairs [article]

Sean T. Yang, Kuan-Hao Huang, Bill Howe
2020 arXiv   pre-print
We propose JECL, a method for clustering image-caption pairs by training parallel encoders with regularized clustering and alignment objectives, simultaneously learning both representations and cluster  ...  JECL trains by minimizing the Kullback-Leibler divergence between the distribution of the images and text to that of a combined joint target distribution and optimizing the Jensen-Shannon divergence between  ...  [14] used a joint matrix factorization with restraints to progressively find the consensus between different views. Zhao et al.  ... 
arXiv:1901.01860v3 fatcat:vqy4c2eejzedzajlxia44t6in4

Cluster-based Joint Matrix Factorization Hashing for Cross-Modal Retrieval

Dimitrios Rafailidis, Fabio Crestani
2016 Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval - SIGIR '16  
In this study, we propose a cross-modal hashing method by following a cluster-based joint matrix factorization strategy.  ...  We formulate a joint matrix factorization process with the constraint that pushes the documents' representations of the different modalities and the cross-modal cluster representations into a common consensus  ...  In addition, images are represented by 128-dimensional SIFT feature vectors [10] and text by 10-dimensional topics vectors. All documents (image-text pairs) are labeled by 10 semantic categories.  ... 
doi:10.1145/2911451.2914710 dblp:conf/sigir/RafailidisC16 fatcat:d6wa7bh5evdktbfmke3fdp4n6a

Wavelet Based Joint Denoising of Depth and Luminance Images

Ljubomir Jovanov, Aleksandra Pizurica, Wilfried Philips
2007 2007 3DTV Conference  
In this paper we present a new method for joint denoising of depth and luminance images produced by time-of-flight camera.  ...  Denoising results are compared with the ground truth images obtained by averaging of the multiple frames of the still scene.  ...  In the case of k-means clustering, centroid values are returned together with labelling.  ... 
doi:10.1109/3dtv.2007.4379389 fatcat:t3zp37k7ybctvfwtv52cfq6zky

Joint NMF for Identification of Shared Features in Datasets and a Dataset Distance Measure [article]

Hannah Friedman, Amani R. Maina-Kilaas, Julianna Schalkwyk, Hina Ahmed, Jamie Haddock
2022 arXiv   pre-print
In this paper, we derive a new method for determining shared features of datasets by employing joint non-negative matrix factorization and analyzing the resulting factorizations.  ...  We also propose a dataset distance measure built upon this method and the learned factorization. Our method is able to successfully identity differences in structure in both image and text datasets.  ...  Joint non-negative matrix factorization (jNMF) allows for joint factorization of two data sets with a common basis [9] , [10] .  ... 
arXiv:2207.05112v1 fatcat:lby6oztm2jhcdppxxmn6b77pge

Robust semi-supervised nonnegative matrix factorization

Jing Wang, Feng Tian, Chang Hong Liu, Xiao Wang
2015 2015 International Joint Conference on Neural Networks (IJCNN)  
Nonnegative matrix factorization (NMF), which aims at finding parts-based representations of nonnegative data, has been widely applied to a wide range of applications such as data clustering, pattern recognition  ...  In this paper, we propose a robust semi-supervised nonnegative matrix factorization (RSSNMF) approach which takes all factors into consideration.  ...  Given a data set containing n images, let l i and r i be the the obtained cluster label and label provided from each sample image, respectively.  ... 
doi:10.1109/ijcnn.2015.7280422 dblp:conf/ijcnn/WangTLW15 fatcat:22ylmrtrkrc6pcymied3zylmqy

Robust Manifold Matrix Factorization for Joint Clustering and Feature Extraction

Lefei Zhang, Qian Zhang, Bo Du, Dacheng Tao, Jane You
) and then perform the matrix approximation guided by such label information.  ...  However, in order to enhance the discriminability, most of the matrix approximation based feature extraction algorithms usually generate the cluster labels by certain clustering algorithm (e.g., the kmeans  ...  Conclusion In this paper, we propose a robust manifold matrix factorization (RMMF) for joint clustering and feature extraction.  ... 
doi:10.1609/aaai.v31i1.10714 fatcat:ca4vzxrq65bwvkiakgarx3i4je

Transfer Nonnegative Matrix Factorization for Image Representation [chapter]

Tianchun Wang, TengQi Ye, Cathal Gurrin
2016 Lecture Notes in Computer Science  
Nonnegative Matrix Factorization (NMF) has received considerable attention due to its psychological and physiological interpretation of naturally occurring data whose representation may be partsbased in  ...  However, when labeled and unlabeled images are sampled from different distributions, they may be quantized into different basis vector space and represented in different coding vector space, which may  ...  Transfer Nonnegative Matrix Factorization In this section, we will present the Transfer Nonnegative Matrix Factorization (TNMF) algorithm for image representation, which extends HeNMF by taking into account  ... 
doi:10.1007/978-3-319-27674-8_1 fatcat:66u6x736ondonh4qy6gw66a2fu

A Bayesian Framework for Modeling Human Evaluations [chapter]

Himabindu Lakkaraju, Jure Leskovec, Jon Kleinberg, Sendhil Mullainathan
2015 Proceedings of the 2015 SIAM International Conference on Data Mining  
Examples of such situations include a crowd-worker labeling an image or a student answering a multiple-choice question.  ...  true labels of items.  ...  factorization.  ... 
doi:10.1137/1.9781611974010.21 dblp:conf/sdm/LakkarajuLKM15 fatcat:37quoraov5ga3i5ms56jkaqw5y

Boundary-aware Information Maximization for Self-supervised Medical Image Segmentation [article]

Jizong Peng, Ping Wang, Marco Pedersoli, Christian Desrosiers
2022 arXiv   pre-print
Among various methods, contrastive learning learns a discriminative representation by constructing positive and negative pairs.  ...  Unsupervised pre-training has been proven as an effective approach to boost various downstream tasks given limited labeled data.  ...  We confirm this by visualizing in Fig. 1 the joint matrix P joint , the cluster assignment, as well as the uncertainty of these clusters for different α.  ... 
arXiv:2202.02371v2 fatcat:sup2l4iahjdl3dajxcytmmmwsq

Automatic annotation of histopathological images using a latent topic model based on non-negative matrix factorization

FabioA González, Gloria Díaz, Angel Cruz-Roa, Eduardo Romero
2011 Journal of Pathology Informatics  
ACKNOWLEDGMENTS This work was partially funded by the projects "Automatic Annotation and Retrieval of Radiology Images Using Latent Semantic" number 110152128803 and "Medical Image Retrieval System Based  ...  Additionally, each latent factor can be associated to a cluster of images, [18] the centroid of the cluster given by the columns of W and the assignment of images to clusters given by the rows of H,  ...  Ding et al. showed that the factorizations produced by NMF and PLSI are equivalent, [16] with W containing the visual-wordlatent-factor conditional probabilities, p(w i |z k ), and H the image-latent-factor  ... 
doi:10.4103/2153-3539.92031 pmid:22811960 pmcid:PMC3312710 fatcat:5hxwlflwsbbudci5launhom2we

Drying of open animal joints in vivo subsequently causes cartilage degeneration

S. I. Paterson, N. M. Eltawil, A. H. R. W. Simpson, A. K. Amin, A. C. Hall
2016 Bone & Joint Research  
By week eight, chondrocyte pairing/clustering and cell volume increased (p < 0.05; p < 0.001, respectively).  ...  Animals were monitored for up to eight weeks and then sacrificed. Cartilage and chondrocyte properties were studied by histology and confocal microscopy, respectively.  ...  By week eight, more CMFDa-labelled chondrocytes were evident Cartilage thickness following in vivo joint drying.  ... 
doi:10.1302/2046-3758.54.2000594 pmid:27114348 pmcid:PMC4921049 fatcat:e6qb4hineredpbzrmov2uvqasq

Dyadic transfer learning for cross-domain image classification

Hua Wang, Feiping Nie, Heng Huang, Chris Ding
2011 2011 International Conference on Computer Vision  
Because manual image annotation is both expensive and labor intensive, in practice we often do not have sufficient labeled images to train an effective classifier for the new image classification tasks  ...  In this paper, we propose a novel nonnegative matrix tri-factorization based transfer learning framework, called as Dyadic Knowledge Transfer (DKT) approach, to transfer cross-domain image knowledge for  ...  This research was supported by NSF-IIS 1117965, NSF-CCF-0830780, NSF-DMS-0915228, NSF-CCF-0917274.  ... 
doi:10.1109/iccv.2011.6126287 dblp:conf/iccv/WangNHD11 fatcat:56i6dffrcfazjmpup62ly6qcau
« Previous Showing results 1 — 15 out of 51,980 results