Filters








48,785 Hits in 3.8 sec

Learning Invariant Visual Shape Representations from Physics [chapter]

Mathias Franzius, Heiko Wersing
2010 Lecture Notes in Computer Science  
The unsupervised bottom-up learning leads to pose invariant representations. Shape specificity is then integrated as top-down information from the movement trajectories of the objects.  ...  In this article, we introduce an autonomous learning system for categorizing 3D shape of simulated objects from single views.  ...  Similar to the results with the additional supervised training views above, this autonomously learned representation categorizes views by object shape.  ... 
doi:10.1007/978-3-642-15825-4_38 fatcat:3kawimjpkjgh3jm2a4ch6vruze

Vision: are models of object recognition catching up with the brain?

Tomaso Poggio, Shimon Ullman
2013 Annals of the New York Academy of Sciences  
and categorization.  ...  Object recognition has been a central yet elusive goal of computational vision.  ...  We will discuss such an approach in the context of learning invariant representations.  ... 
doi:10.1111/nyas.12148 pmid:23773126 fatcat:j3t3es4iwvauvl5jxy2dv5zasi

Mechanisms of object recognition: what we have learned from pigeons

Fabian A. Soto, Edward A. Wasserman
2014 Frontiers in Neural Circuits  
This possibility has remained unexplored in the primate literature, which has focused instead on looking for evidence of unsupervised learning of invariant object representations (Cox et al., 2005; Li  ...  instead of visual object representation, which is different from the focus of most human research (Soto and Wasserman, 2012b) .  ...  Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest  ... 
doi:10.3389/fncir.2014.00122 pmid:25352784 pmcid:PMC4195317 fatcat:y7hzfh3555akddffcuan6imxty

CBF: A New Framework for Object Categorization in Cortex [chapter]

Maximilian Riesenhuber, Tomaso Poggio
2000 Lecture Notes in Computer Science  
We demonstrate the capability of our scheme, called "Categorical Basis Functions" (CBF) with the example domain of cat/dog categorization, using stimuli generated with a novel 3D morphing system.  ...  Building on our recent hierarchical model of object recognition in cortex, we show how this model can be extended in a straightforward fashion to perform basic-level object categorization.  ...  Thanks to Christian Shelton for MATLAB code for k-means and RBF training and for the development of the correspondence program [17] .  ... 
doi:10.1007/3-540-45482-9_1 fatcat:d5qqqef2yfhd3gpmbva2wujk5u

Categorization training changes the visual representation of face identity

Fabian A. Soto
2019 Attention, Perception & Psychophysics  
Previous research suggests that learning to categorize faces along a new dimension changes the perceptual representation of that dimension, but little is known about how the representation of specific  ...  These results suggest that the representation of face identity can be modified by categorization experience.  ...  The work described here conformed to Standard 8 of the American Psychological Association's Ethical principles of Psychologists and Code of Conduct (http://www.apa.org/ethics/code/ index.aspx).  ... 
doi:10.3758/s13414-019-01765-w pmid:31152373 pmcid:PMC6703175 fatcat:jdik7lur3vbvnoa3oo35lpuqsq

MiniMax Entropy Network: Learning Category-Invariant Features for Domain Adaptation [article]

Chaofan Tao, Fengmao Lv, Lixin Duan, Min Wu
2019 arXiv   pre-print
Unlike most existing approaches which employ a generator to deal with domain difference, MMEN focuses on learning the categorical information from unlabeled target samples with the help of labeled source  ...  Specifically, we set an unfair multi-class classifier named categorical discriminator, which classifies source samples accurately but be confused about the categories of target samples.  ...  ., 2017; Chadha and Andreopoulos, 2018] primarily focused on learning domain-invariant representations, by which the distribution discrepancy between domains can be reduced.  ... 
arXiv:1904.09601v2 fatcat:3jo5pzlegff25epilartkvzhem

Diagnostic recognition: task constraints, object information, and their interactions

Philippe G Schyns
1998 Cognition  
It is the main objective of this paper to lay out the basis of a dialogue between object recognition and categorization research, with the hope of raising issues that could cross-fertilize both domains  ...  Object recognition and categorization research are both concerned with understanding how input information matches object information in memory.  ...  Acknowledgements The author wishes to thank Michael Burton, Simon Garrod, Frédéric Gosselin, Gregory Murphy, Martin Pickering and two anonymous reviewers for helpful comments on an earlier version of this  ... 
doi:10.1016/s0010-0277(98)00016-x pmid:9735539 fatcat:2smkqkrkhbgfxmpwaolhnfun3i

Models of object recognition

Maximilian Riesenhuber, Tomaso Poggio
2000 Nature Neuroscience  
From the computational viewpoint of learning, different recognition tasks, such as categorization and identification, are similar, representing different trade-offs between specificity and invariance.  ...  Thus, the different tasks do not require different classes of models.  ...  Schemes for learning overcomplete representations have been proposed 56 , with extensions to the learning of invariances 57 .  ... 
doi:10.1038/81479 pmid:11127838 fatcat:jj5g4uuodbfjfl5aokthavvqv4

Are we done with object recognition? The iCub robot's perspective

Giulia Pasquale, Carlo Ciliberto, Francesca Odone, Lorenzo Rosasco, Lorenzo Natale
2019 Robotics and Autonomous Systems  
We report on an extensive study of the benefits and limitations of current deep learning approaches to object recognition in robot vision scenarios, introducing a novel dataset used for our investigation  ...  deep learning approaches have been originally designed.  ...  The most critical parameters were instead the base LR and the numbers of FC layers learned from scratch.  ... 
doi:10.1016/j.robot.2018.11.001 fatcat:tjmnbom4ungr7iketpff3kz52i

From Sensations to Concepts: a Proposal for Two Learning Processes

Peter Gärdenfors
2018 Review of Philosophy and Psychology  
The invariant structures involve a reduction in dimensionality of the sensory information.  ...  The first process constructs the perceptual structures that emerge in children's cognitive development by detecting invariants in the sensory input.  ...  the workshop on Concept Learning and Reasoning in Conceptual Spaces in Bochum for helpful comments on earlier versions of this paper.  ... 
doi:10.1007/s13164-017-0379-7 fatcat:rjejn4mhhjd25lrw4jotrdqwum

Learning-induced categorical perception in a neural network model [article]

Christian Thériault, Fernanda Pérez-Gay, Dan Rivas, Stevan Harnad
2018 arXiv   pre-print
In human cognition, the expansion of perceived between-category distances and compression of within-category distances is known as categorical perception (CP).  ...  There are several hypotheses about the causes of CP (e.g., language, learning, evolution) but no functional model.  ...  In learning the conjunctive rule of Simulation 1, the spatial invariance properties of convolutional networks (a network composed of local filter providing invariance to position) would perform better  ... 
arXiv:1805.04567v1 fatcat:yv2pca7uobefnlc6hrhisf2n5e

Unsupervised Learning of Individuals and Categories from Images

Stephen Waydo, Christof Koch
2008 Neural Computation  
In our model, a network of nonlinear neurons learns a sparse representation of its inputs through an unsupervised expectationmaximization process.  ...  We show that the application of this strategy to an invariant feature-based description of natural images leads to the development of units displaying sparse, invariant selectivity for particular individuals  ...  Thomas Serre and Minjoon Kouh of MIT provided invaluable assistance in the setup and operation of the underlying vision model.  ... 
doi:10.1162/neco.2008.03-07-493 fatcat:i6rv5g7bkfcnhnz7rycg2zolcy

Unsupervised Learning of Individuals and Categories from Images

Stephen Waydo, Christof Koch
2008 Neural Computation  
In our model, a network of nonlinear neurons learns a sparse representation of its inputs through an unsupervised expectationmaximization process.  ...  We show that the application of this strategy to an invariant feature-based description of natural images leads to the development of units displaying sparse, invariant selectivity for particular individuals  ...  Thomas Serre and Minjoon Kouh of MIT provided invaluable assistance in the setup and operation of the underlying vision model.  ... 
doi:10.1162/neco.2007.03-07-493 pmid:18194101 fatcat:4m5woeqzrfhfhenfxym4i2dgmy

Interleaving Object Categorization and Segmentation [chapter]

Bastian Leibe, Bernt Schiele
2006 Lecture Notes in Computer Science  
Schiele, CVPR'05 Interleaved Object Categorization and Segmentation Codebook Representation • Extraction of local object patches Interest Points (e.g.  ...  Introduction Interleaved Object Categorization and Segmentation Learning the Spatial Layout ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) x o p , p I p I x o p x o p , p , x o p x o , • Approach Interleaved categorization  ... 
doi:10.1007/11414353_10 fatcat:t5wmvaxqmjflnl7jy46p2grahq

Integrating Categorical Semantics into Unsupervised Domain Translation [article]

Samuel Lavoie, Faruk Ahmed, Aaron Courville
2021 arXiv   pre-print
We propose a method to learn, in an unsupervised manner, categorical semantic features (such as object labels) that are invariant of the source and target domains.  ...  We show that conditioning the style encoder of unsupervised domain translation methods on the learned categorical semantics leads to a translation preserving the digits on MNIST↔SVHN and to a more realistic  ...  We acknowledge financial support of Hitachi, Samsung, CIFAR and the Natural Sciences and Engineering Research Council of Canada (NSERC Discovery Grant).  ... 
arXiv:2010.01262v2 fatcat:4762qyrt5jhtrlcd2m3etgtghe
« Previous Showing results 1 — 15 out of 48,785 results