Book contents
- Frontmatter
- Contents
- Preface
- Contributors
- 1 The Evolution of Object Categorization and the Challenge of Image Abstraction
- 2 A Strategy for Understanding How the Brain Accomplishes Object Recognition
- 3 Visual Recognition Circa 2008
- 4 On What It Means to See, and WhatWe Can Do About It
- 5 Generic Object Recognition by Inference of 3-D Volumetric Parts
- 6 What Has fMRI Taught Us About Object Recognition?
- 7 Object Recognition Through Reasoning About Functionality: A Survey of Related Work
- 8 The Interface Theory of Perception: Natural Selection Drives True Perception to Swift Extinction
- 9 Words and Pictures: Categories, Modifiers, Depiction, and Iconography
- 10 Structural Representation of Object Shape in the Brain
- 11 Learning Hierarchical Compositional Representations of Object Structure
- 12 Object Categorization in Man, Monkey, and Machine: Some Answers and Some Open Questions
- 13 Learning Compositional Models for Object Categories from Small Sample Sets
- 14 The Neurophysiology and Computational Mechanisms of Object Representation
- 15 From Classification to Full Object Interpretation
- 16 Visual Object Discovery
- 17 Towards Integration of Different Paradigms in Modeling, Representation, and Learning of Visual Categories
- 18 Acquisition and Disruption of Category Specificity in the Ventral Visual Stream: The Case of Late Developing and Vulnerable Face-Related Cortex
- 19 Using Simple Features and Relations
- 20 The Proactive Brain: Using Memory-Based Predictions in Visual Recognition
- 21 Spatial Pyramid Matching
- 22 Visual Learning for Optimal Decisions in the Human Brain
- 23 Shapes and Shock Graphs: From Segmented Shapes to Shapes Embedded in Images
- 24 Neural Encoding of Scene Statistics for Surface and Object Inference
- 25 Medial Models for Vision
- 26 Multimodal Categorization
- 27 Comparing 2-D Images of 3-D Objects
- Index
- Plate section
17 - Towards Integration of Different Paradigms in Modeling, Representation, and Learning of Visual Categories
Published online by Cambridge University Press: 20 May 2010
- Frontmatter
- Contents
- Preface
- Contributors
- 1 The Evolution of Object Categorization and the Challenge of Image Abstraction
- 2 A Strategy for Understanding How the Brain Accomplishes Object Recognition
- 3 Visual Recognition Circa 2008
- 4 On What It Means to See, and WhatWe Can Do About It
- 5 Generic Object Recognition by Inference of 3-D Volumetric Parts
- 6 What Has fMRI Taught Us About Object Recognition?
- 7 Object Recognition Through Reasoning About Functionality: A Survey of Related Work
- 8 The Interface Theory of Perception: Natural Selection Drives True Perception to Swift Extinction
- 9 Words and Pictures: Categories, Modifiers, Depiction, and Iconography
- 10 Structural Representation of Object Shape in the Brain
- 11 Learning Hierarchical Compositional Representations of Object Structure
- 12 Object Categorization in Man, Monkey, and Machine: Some Answers and Some Open Questions
- 13 Learning Compositional Models for Object Categories from Small Sample Sets
- 14 The Neurophysiology and Computational Mechanisms of Object Representation
- 15 From Classification to Full Object Interpretation
- 16 Visual Object Discovery
- 17 Towards Integration of Different Paradigms in Modeling, Representation, and Learning of Visual Categories
- 18 Acquisition and Disruption of Category Specificity in the Ventral Visual Stream: The Case of Late Developing and Vulnerable Face-Related Cortex
- 19 Using Simple Features and Relations
- 20 The Proactive Brain: Using Memory-Based Predictions in Visual Recognition
- 21 Spatial Pyramid Matching
- 22 Visual Learning for Optimal Decisions in the Human Brain
- 23 Shapes and Shock Graphs: From Segmented Shapes to Shapes Embedded in Images
- 24 Neural Encoding of Scene Statistics for Surface and Object Inference
- 25 Medial Models for Vision
- 26 Multimodal Categorization
- 27 Comparing 2-D Images of 3-D Objects
- Index
- Plate section
Summary
Introduction
Object representations for categorization tasks should be applicable for a wide range of objects, scalable to handle large numbers of object classes, and at the same time learnable from a few training samples. While such a scalable representation is still illusive today, it has been argued that such a representation should have at least the following properties: it should enable sharing of features (Torralba et al. 2007), it should combine generative models with discriminative models (Fritz et al. 2005; Jaakkola and Haussler 1999), and it should combine both local and global as well as appearanceand shape-based features (Leibe et al. 2005). Additionally, we argue that such object representations should be applicable both for unsupervised learning (e.g., visual object discovery) as well as supervised training (e.g., object detection). Therefore, we extend our previous efforts of hybrid modeling (Fritz et al. 2005) with ideas of unsupervised learning of generative decompositions to obtain an approach that integrates across different paradigms of modeling, representing, and learning of visual categories.
We present a novel method for the discovery and detection of visual object categories based on decompositions using topic models. The approach is capable of learning a compact and low-dimensional representation for multiple visual categories from multiple viewpoints without labeling of training instances. The learnt object components range from local structures over line segments to global silhouette-like descriptions. This representation can be used to discover object categories in a totally unsupervised fashion. Furthermore we employ the representation as the basis for building a supervised multicategory detection system making efficient use of training examples and outperforming pure features-based representations.
- Type
- Chapter
- Information
- Object CategorizationComputer and Human Vision Perspectives, pp. 324 - 347Publisher: Cambridge University PressPrint publication year: 2009