4,253 Hits in 7.0 sec

Mutual Information Independence Model Using Kernel Density Estimation for Segmenting and Labeling Sequential Data [chapter]

Guodong Zhou, Lingpeng Yang, Jian Su, Donghong Ji
2005 Lecture Notes in Computer Science  
This paper proposes a Mutual Information Independence Model (MIIM) to segment and label sequential data.  ...  In addition, a variable-length pairwise mutual information-based modeling approach and a kNN algorithm using kernel density estimation are proposed to capture the long state dependence and the observation  ...  Given an observation sequence (1) Traditionally, HMM segments and labels sequential data in a generative way by making a context independent assumption that successive observations are independent given  ... 
doi:10.1007/978-3-540-30586-6_15 fatcat:okhxdsjhtnfdpnkwm75gfb43le

Visual Cue Cluster Construction via Information Bottleneck Principle and Kernel Density Estimation [chapter]

Winston H. Hsu, Shih-Fu Chang
2005 Lecture Notes in Computer Science  
., color, texture, motion, etc.) and used for discriminative classification of semantic labels. However, in most systems, such mid-level features are selected manually.  ...  The problem is posed as mutual information maximization, through which optimal cue clusters are discovered to preserve the highest information about the semantic labels.  ...  Acknowledgments We thank Dan Ellis and Lyndon Kennedy of Columbia University for useful discussions and V. France of [12] for his kind support of LDA implementation.  ... 
doi:10.1007/11526346_12 fatcat:zki76rdqmzaypdul5ebm6wgnhe

Sequential Convex Relaxation for Mutual Information-Based Unsupervised Figure-Ground Segmentation

Youngwook Kee, Mohamed Souiai, Daniel Cremers, Junmo Kim
2014 2014 IEEE Conference on Computer Vision and Pattern Recognition  
To this end, we revisit the notion of mutual information and reformulate it in terms of the photometric variable and the indicator function; and propose a sequential convex optimization strategy for solving  ...  The algorithm jointly estimates the color distributions of the foreground and background, and separates them based on their mutual information with geometric regularity.  ...  For the step sizes (τ , σ), we chose sufficiently small values. Kernel Density Estimation We use a Gaussian function with a diagonal matrix H for K H (z).  ... 
doi:10.1109/cvpr.2014.520 dblp:conf/cvpr/KeeSCK14 fatcat:bmu2grwehrfgrgbk4srzkkd5oa

Kernel Measures of Independence for non-iid Data

Xinhua Zhang, Le Song, Arthur Gretton, Alexander J. Smola
2008 Neural Information Processing Systems  
This is achieved by modeling the structures using undirected graphical models and comparing the Hilbert space embeddings of distributions.  ...  We apply this new criterion to independent component analysis and sequence clustering.  ...  Acknowledgements NICTA is funded by the Australian Governments Backing Australia's Ability and the Centre of Excellence programs.  ... 
dblp:conf/nips/ZhangSGS08 fatcat:r2bpiajmkbgkhcuqfmoik7wsre

Information criteria performance for feature selection

Mohamed Abadi, Olivier Alata, Christian Olivier, Majdi Khoudeir, Enguerran Grandchamp
2011 2011 4th International Congress on Image and Signal Processing  
Information Criteria could be used for feature selection as a good alternative to other criteria.  ...  We apply this approach to classify an hand segmented image. The performance is tested using various feature selection schemes (SFS, SBS, SFFS and SBFS) to select the candidate subsets.  ...  ACKNOWLEDGMENT The authors want to thanks the European institutions and more precisely the INTERREG IV program within the CESAR Part II project leads by the University of Antilles and Guyana and the Poitou-Charentes  ... 
doi:10.1109/cisp.2011.6100275 fatcat:fjyugikfxfhdhccc62st7mowtq

Outlier Correction for Local Distance Measures in Example Based Speech Recognition

Mathias De Wachter, Kris Demuynck, Dirk Van Compernolle
2007 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07  
Experiments on three low-level acoustic tasks show that "data sharpening" results in a substantial improvement, while "adaptive kernels" have minimal effect.  ...  In this paper, we derive two techniques inspired by non-parametric density estimation that explicitly adjust the distance measure based on the position of the reference vector in its class.  ...  Choi and Hall [11] used the same idea as a preprocessing step for non-parametric density estimation, calling it data sharpening.  ... 
doi:10.1109/icassp.2007.366942 dblp:conf/icassp/WachterDC07 fatcat:oqbl577g4rafjbwzbez35ogsnm

A Multichannel Markov Random Field Framework for Tumor Segmentation With an Application to Classification of Gene Expression-Based Breast Cancer Recurrence Risk

Ahmed B. Ashraf, Sara C. Gavenonis, Dania Daye, Carolyn Mies, Mark A. Rosen, Despina Kontos
2013 IEEE Transactions on Medical Imaging  
We use conditional mutual information to search for features that satisfy conditional independence assumptions.  ...  with high and low recurrence risk, giving an AUC of 0.88 as compared to 0.79, 0.76, 0.75, and 0.66 when using normalized cuts, single channel MRF, FCM, and FCM-VES, respectively, for segmentation.  ...  from the National Center for Research Resources.  ... 
doi:10.1109/tmi.2012.2219589 pmid:23008246 pmcid:PMC4197832 fatcat:6hsllbhgy5butcbes5sfyz7osu

Bayesian Active Meta-Learning for Black-Box Optimization [article]

Ivana Nikoloska, Osvaldo Simeone
2022 arXiv   pre-print
To decrease the number of labeling steps required for meta-learning, this paper introduces an information-theoretic active task selection mechanism, and evaluates an instantiation of the approach for Bayesian  ...  In practice, one may have available only unlabeled data sets from the related tasks, requiring a costly labeling procedure to be carried out before use in meta-learning.  ...  scheme for which the mean and kernel are meta-optimized in the offline phase using all |T a | = 20 metatraining data sets [14] .  ... 
arXiv:2110.09943v2 fatcat:kh5k3m7v2nf3vnkvu243zldwz4

Machine learning and radiology

Shijun Wang, Ronald M. Summers
2012 Medical Image Analysis  
diagnosis from fMR images, content-based image retrieval systems for CT or MRI images, and text analysis of radiology reports using natural language processing (NLP) and natural language understanding  ...  Machine learning identifies complex patterns automatically and helps radiologists make intelligent decisions on radiology data such as conventional radiographs, CT, MRI, and PET images and radiology reports  ...  Acknowledgments We thank Andrew Dwyer, MD, for critical review of the manuscript. This manuscript was support by the Intramural Research Program of the National Institutes of Health Clinical Center.  ... 
doi:10.1016/ pmid:22465077 pmcid:PMC3372692 fatcat:4ynexgzdhrev7dfqapmjpxexuu

An unsupervised data projection that preserves the cluster structure

Lev Faivishevsky, Jacob Goldberger
2012 Pattern Recognition Letters  
Instead, we utilize a non-parametric estimation of the average cluster entropies and search for a linear projection and a clustering that maximizes the estimated mutual information between the projected  ...  In order to compute the mutual information, we neither assume the data are given in terms of distributions nor impose any parametric model on the within-cluster distribution.  ...  In addition, the Kernel methods may be applicable for the proposed framework of simultaneous dimensionality reduction and clustering.  ... 
doi:10.1016/j.patrec.2011.10.012 fatcat:aiefszh22rgnlnzuu6hks5wf5q

Nonparametric Information Theoretic Clustering Algorithm

Lev Faivishevsky, Jacob Goldberger
2010 International Conference on Machine Learning  
Instead, we utilize a non-parametric estimation of the average cluster entropies and search for a clustering that maximizes the estimated mutual information between data points and clusters.  ...  In this paper we propose a novel clustering algorithm based on maximizing the mutual information between data points and clusters.  ...  In particular, we maximize the mutual information between cluster labels and features of data points without imposing any parametric model on the cluster distribution.  ... 
dblp:conf/icml/FaivishevskyG10 fatcat:glgnovmgrbhq3j7tckzwregbx4

Machine Learning Techniques for the Analysis of Magnetic Flux Leakage Images in Pipeline Inspection

A. Khodayari-Rostamabad, J.P. Reilly, N.K. Nikolova, J.R. Hare, S. Pasha
2009 IEEE transactions on magnetics  
We demonstrate the adequacy of the performance of these methods using real MFL data collected from pipelines, with regard to the performance of both the detection of defects, and the accuracy in the estimation  ...  The magnetic flux leakage (MFL) technique is commonly used for non-destructive testing of oil and gas pipelines.  ...  The authors would like to thank Intratech Inline Inspection Services Ltd., and specifically its president, Ron Thompson, for supporting this research and providing the required MFL data.  ... 
doi:10.1109/tmag.2009.2020160 fatcat:ynpu6xitk5ag3omv4mowbaxjla

Efficient Monte Carlo Sampler for Detecting Parametric Objects in Large Scenes [chapter]

Yannick Verdié, Florent Lafarge
2012 Lecture Notes in Computer Science  
However, simulating these mathematical models is a difficult task, especially on large scenes. Existing samplers suffer from average performances in terms of computation time and stability.  ...  This procedure is embedded into a data-driven mechanism such that the points are non-uniformly distributed in the scene.  ...  The authors thank the French Mapping Agency (IGN) and the Tour du Valat for provinding the datasets.  ... 
doi:10.1007/978-3-642-33712-3_39 fatcat:tfd25glhzrd6bcgfmpzenbmqeu

A hybrid deep learning framework for integrated segmentation and registration: evaluation on longitudinal white matter tract changes [article]

Bo Li, Wiro Niessen, Stefan Klein, Marius de Groot, Arfan Ikram, Meike Vernooij, Esther Bron
2019 arXiv   pre-print
Existing solutions often involve independent registration and segmentation components.  ...  Registration between time-points is used either as a prior for segmentation in a subsequent time point or to perform segmentation in a common space.  ...  Discussion and conclusion We propose a novel hybrid deep learning framework for integrated segmentation and deformable registration in a single fast procedure.  ... 
arXiv:1908.10221v1 fatcat:vphuywjrhndflj7vopd7szsl74

Semisupervised Hyperspectral Image Segmentation Using Multinomial Logistic Regression With Active Learning

Jun Li, José M. Bioucas-Dias, Antonio Plaza
2010 IEEE Transactions on Geoscience and Remote Sensing  
of jointly considering spatial and spectral information in hyperspectral image segmentation.  ...  The posterior class distributions are modeled using multinomial logistic regression, where the regressors are learned using both labeled and, through a graph-based technique, unlabeled samples.  ...  Gualtieri for providing the AVIRIS Salinas data set used in our experiments.  ... 
doi:10.1109/tgrs.2010.2060550 fatcat:u35nd3zilfhfdb2xicmvepx2by
« Previous Showing results 1 — 15 out of 4,253 results