Filters








2,927 Hits in 3.2 sec

Unsupervised image embedding using nonparametric statistics

Guobiao Mei, Christian R. Shelton
2008 Pattern Recognition (ICPR), Proceedings of the International Conference on  
We make no neighborhood assumptions in our embedding. Our algorithm can also embed the images in a discrete grid, useful for many visualization tasks.  ...  Embedding images into a low dimensional space has a wide range of applications: visualization, clustering, and pre-processing for supervised learning.  ...  In addition, we propose to use the nonparametric statistic Kendall's τ [3] as a criterion to evaluate the embedding quality. In this work, we adopt the overall structure of [5] .  ... 
doi:10.1109/icpr.2008.4761051 dblp:conf/icpr/MeiS08 fatcat:6cq2enmbjne5lbt3ki7whrupha

Unsupervised Texture Segmentation with Nonparametric Neighborhood Statistics [chapter]

Suyash P. Awate, Tolga Tasdizen, Ross T. Whitaker
2006 Lecture Notes in Computer Science  
This paper presents a novel approach to unsupervised texture segmentation that relies on a very general nonparametric statistical model of image neighborhoods.  ...  This paper presents a novel approach to unsupervised texture segmentation that relies on a very general nonparametric statistical model of image neighborhoods.  ...  Moran Eye Center, University of Utah for providing the electron-microscopy retinal images.  ... 
doi:10.1007/11744047_38 fatcat:5eseppz44rckfomznvnme7hn4q

Meibography Phenotyping and Classification From Unsupervised Discriminative Feature Learning

Chun-Hsiao Yeh, Stella X. Yu, Meng C. Lin
2021 Translational Vision Science & Technology  
Four hundred ninety-seven meibography images were used for network learning and tuning, whereas the remaining 209 images were used for network model evaluations.  ...  One of the latest unsupervised learning approaches is to apply feature learning based on nonparametric instance discrimination (NPID), a convolutional neural network (CNN) backbone model trained to encode  ...  The authors thank Dorothy Ng, Jessica Vu, Jasper Cheng, Kristin Kiang, Megan Tsiu, Fozia KhanRam, April Myers, Shawn Tran, Michelle Hoang, and Zoya Razzak for providing annotations for the meibography images  ... 
doi:10.1167/tvst.10.2.4 pmid:34003889 pmcid:PMC7873493 fatcat:7rc6mabqyfhxfgg2m7d4j4t7va

DIMENSIONALITY REDUCTION OF HYPERSPECTRAL IMAGES BY COMBINATION OF NON-PARAMETRIC WEIGHTED FEATURE EXTRACTION (NWFE) AND MODIFIED NEIGHBORHOOD PRESERVING EMBEDDING (NPE)

T. Alipour Fard, H. Arefi
2014 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
NPE Neighborhood Preserving Embedding is a linear approximation to locally linear embedding feature extraction.  ...  Neighborhood Preserving Embedding (NPE) is another unsupervised method can preserve local neighborhood information and overcome to over fitting problems of supervised methods (Liao et al, 2011) .  ... 
doi:10.5194/isprsarchives-xl-2-w3-31-2014 fatcat:mcsx75w47bhm7o5i2gvkd2xjoe

FINE: Fisher Information Nonparametric Embedding

K.M. Carter, R. Raich, W.G. Finn, A.O. Hero
2009 IEEE Transactions on Pattern Analysis and Machine Intelligence  
As a whole, we refer to our framework as Fisher Information Nonparametric Embedding (FINE) and illustrate its uses on practical problems, including a biomedical application and document classification.  ...  In this paper, we propose using the properties of information geometry and statistical manifolds in order to define similarities between data sets using the Fisher information distance.  ...  This work is partially funded by US National Science Foundation, grant No. CCR-0325571.  ... 
doi:10.1109/tpami.2009.67 pmid:19762935 fatcat:3oyjjiyohncb3psuuamqmexuhq

Information Preserving Embeddings for Discrimination

Kevin M. Carter, Christine Kyung-min Kim, Raviv Raich, Alfred O. Hero
2009 2009 IEEE 13th Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop  
Using Fisher information nonparametric embedding, we were able to reconstruct the statistical manifold in an low-dimensional Euclidean space.  ...  We also employ the use of Fisher information nonparametric embedding (FINE) [2, 3] , which provides an information-geometric embedding of the image sets into a low-dimensional Euclidean space.  ... 
doi:10.1109/dsp.2009.4785953 fatcat:jlba2mqnd5ecngkg7a6677pjgi

A Kernel Approach to Tractable Bayesian Nonparametrics [article]

Ferenc Huszár, Simon Lacoste-Julien
2011 arXiv   pre-print
Inference in popular nonparametric Bayesian models typically relies on sampling or other approximations.  ...  This paper presents a general methodology for constructing novel tractable nonparametric Bayesian methods by applying the kernel trick to inference in a parametric Bayesian model.  ...  Nevertheless, the unnormalised density can still be used in a variety of unsupervised learning tasks, such as image reconstruction or novelty detection.  ... 
arXiv:1103.1761v3 fatcat:77rkxdijdzbj7enx4nxwe2rr54

Rethinking Semantic Segmentation: A Prototype View [article]

Tianfei Zhou, Wenguan Wang, Ender Konukoglu, Luc Van Gool
2022 arXiv   pre-print
This allows our model to directly shape the pixel embedding space, by optimizing the arrangement between embedded pixels and anchored prototypes.  ...  The dense prediction is thus achieved by nonparametric nearest prototype retrieving.  ...  By sharing such regime, our nonparametric model has good potential to make full use of unsupervised representations. • Further Enhancing Interpretability.  ... 
arXiv:2203.15102v2 fatcat:hlbuwxv5mnejzaqwg6bncbxski

A Bayesian Treatment of Real-to-Sim for Deformable Object Manipulation

Rika Antonova, Jingyun Yang, Priya Sundaresan, Dieter Fox, Fabio Ramos, Jeannette Bohg
2022 IEEE Robotics and Automation Letters  
A sequence of images of a moving object can then be represented by a trajectory of distribution embeddings in this novel state space for deformables.  ...  Despite only using a small set of real-world trajectories, we show how the proposed approach can estimate posterior distributions over simulation parameters, such as elasticity, friction and scale, even  ...  =1 ) to compute the statistics. [31] applied BayesSim to scenarios with granular media and developed domain-specific summary statistics of depth images of granular formations, such as dispersion and statistical  ... 
doi:10.1109/lra.2022.3157377 fatcat:iwvginpvuvb2lfwbjajxosr65e

An Automatic Method for Unsupervised Feature Selection of Hyperspectral Images Based on Fuzzy Clustering of Bands

Behnam Beirami, Mehdi Mokhtarzade
2020 Traitement du signal  
In this study, an automatic unsupervised method is presented for feature selection from hyperspectral images.  ...  This feature space is originated from the statistical attributes of image bands while these attributes are extracted from different partitions of the entire image.  ...  METHODOLOGY As mentioned earlier, MAD and kurtosis measures are used as statistical measures [32] . It seems that using the more statistical measures can enhanced their method.  ... 
doi:10.18280/ts.370218 fatcat:x2qqt2vdpzd5parrjfpubfhb2m

Learning Robust Visual-Semantic Embeddings [article]

Yao-Hung Hubert Tsai and Liang-Kang Huang and Ruslan Salakhutdinov
2017 arXiv   pre-print
Many of the existing methods for learning joint embedding of images and text use only supervised information from paired images and its textual attributes.  ...  A novel technique of unsupervised-data adaptation inference is introduced to construct more comprehensive embeddings for both labeled and unlabeled data.  ...  Table 8 lists the statistics of β and λ. Next, we study the power of unsupervised information.  ... 
arXiv:1703.05908v2 fatcat:bmvr3bbvavepbg7k7gwgks7gne

Unsupervised nonparametric detection of unknown objects in noisy images based on percolation theory [article]

Mikhail A. Langovoy and Olaf Wittich and Patrick Laurie Davies
2018 arXiv   pre-print
We develop an unsupervised, nonparametric, and scalable statistical learning method for detection of unknown objects in noisy images.  ...  The method uses results from percolation theory and random graph theory.  ...  Main contributions From the statistical point of view, we treat the object detection problem as a nonparametric hypothesis testing problem within the class of statistical inverse problems on networks.  ... 
arXiv:1102.5019v2 fatcat:eahqbco5nravtgygcddwnyrqga

Bayesian Nonparametric Clustering for Positive Definite Matrices

Anoop Cherian, Vassilios Morellas, Nikolaos Papanikolopoulos
2016 IEEE Transactions on Pattern Analysis and Machine Intelligence  
To address this issue, we resort to the classical nonparametric Bayesian framework by modeling the data as a mixture model using the Dirichlet process (DP) prior.  ...  Symmetric Positive Definite (SPD) matrices emerge as data descriptors in several applications of computer vision such as object tracking, texture recognition, and diffusion tensor imaging.  ...  Nonparametric clustering of SPD objects via embedding them in the Euclidean space is considered in [26] .  ... 
doi:10.1109/tpami.2015.2456903 pmid:27046838 fatcat:rsclhggddrdd3nvb3xbm7r7pjq

Unseen Object Instance Segmentation with Fully Test-time RGB-D Embeddings Adaptation [article]

Lu Zhang, Siqi Zhang, Xu Yang, Zhiyong Liu
2022 arXiv   pre-print
Specifically, we propose a framework to conduct the Fully Test-time RGB-D Embeddings Adaptation (FTEA) based on parameters of the BatchNorm layer.  ...  The proposed method can be efficiently conducted with test-time images, without requiring annotations or revisiting the large-scale synthetic training data.  ...  We use the UCN [2] as our pre-trained model due to its conciseness and end-to-end fashion. 2) The Pipeline: During test time, we construct a nonparametric entropy objective (NEO) with the unsupervised  ... 
arXiv:2204.09847v1 fatcat:wsskwn6ypfaczd4beoamni3sjq

Unsupervised nonlinear dimensionality reduction machine learning methods applied to multiparametric MRI in cerebral ischemia: preliminary results

Vishwa S. Parekh, Jeremy R. Jacobs, Michael A. Jacobs, Sebastien Ourselin, Martin A. Styner
2014 Medical Imaging 2014: Image Processing  
NLDR methods are a class of algorithms that uses mathematically defined manifolds for statistical sampling of multidimensional classes to generate a discrimination rule of guaranteed statistical accuracy  ...  The evaluation and treatment of acute cerebral ischemia requires a technique that can determine the total area of tissue at risk for infarction using diagnostic magnetic resonance imaging (MRI) sequences  ...  an embedded image [8] .  ... 
doi:10.1117/12.2044001 dblp:conf/miip/ParekhJJ14 fatcat:agizlbaenbfgllrbvfuank3xgu
« Previous Showing results 1 — 15 out of 2,927 results