Visual Object Recognition

✍ Scribed by Kristen Grauman, Bastian Leibe

• Publisher: Springer
• Year: 2011
• Tongue: English
• Leaves: 172
• Series: Synthesis Lectures on Artificial Intelligence and Machine Learning
• Edition: 1
• Category: Library

✦ Synopsis

The visual recognition problem is central to computer vision research. From robotics to information retrieval, many desired applications demand the ability to identify and localize categories, places, and objects. This tutorial overviews computer vision algorithms for visual object recognition and image classification. We introduce primary representations and learning approaches, with an emphasis on recent advances in the field. The target audience consists of researchers or students working in AI, robotics, or vision who would like to understand what methods and representations are available for these problems. This lecture summarizes what is and isn't possible to do reliably today, and overviews key concepts that could be employed in systems requiring visual categorization. Table of Contents: Introduction / Overview: Recognition of Specific Objects / Local Features: Detection and Description / Matching Local Features / Geometric Verification of Matched Features / Example Systems: Specific-Object Recognition / Overview: Recognition of Generic Object Categories / Representations for Object Categories / Generic Object Detection: Finding and Scoring Candidates / Learning Generic Object Category Models / Example Systems: Generic Object Recognition / Other Considerations and Current Challenges / Conclusions

✦ Table of Contents

Cover
Copyright Page
Title Page
Contents
Preface
Acknowledgments
Figure Credits
Introduction
Overview
Challenges
The State of the Art
Overview: Recognition of Specific Objects
Global Image Representations
Local Feature Representations
Local Features: Detection and Description
Introduction
Detection of Interest Points and Regions
Keypoint Localization
Scale Invariant Region Detection
Affine Covariant Region Detection
Orientation Normalization
Summary of Local Detectors
Local Descriptors
The SIFT Descriptor
The SURF Detector/Descriptor
Concluding Remarks
Matching Local Features
Efficient Similarity Search
Tree-based Algorithms
Hashing-based Algorithms and Binary Codes
A Rule of Thumb for Reducing Ambiguous Matches
Indexing Features with Visual Vocabularies
Creating a Visual Vocabulary
Vocabulary Trees
Choices in Vocabulary Formation
Inverted File Indexing
Concluding Remarks
Geometric Verification of Matched Features
Estimating Geometric Models
Estimating Similarity Transformations
Estimating Affine Transformations
Homography Estimation
More General Transformations
Dealing with Outliers
RANSAC
Generalized Hough Transform
Discussion
Example Systems: Specific-Object Recognition
Image Matching
Object Recognition
Large-Scale Image Retrieval
Mobile Visual Search
Image Auto-Annotation
Concluding Remarks
Overview: Recognition of Generic Object Categories
Representations for Object Categories
Window-based Object Representations
Pixel Intensities and Colors
Window Descriptors: Global Gradients and Texture
Patch Descriptors: Local Gradients and Texture
A Hybrid Representation: Bags of Visual Words
Contour and Shape Features
Feature Selection
Part-based Object Representations
Overview of Part-Based Models
Fully-Connected Models: The Constellation Model
Star Graph Models
Mixed Representations
Concluding Remarks
Generic Object Detection: Finding and Scoring Candidates
Detection via Classification
Speeding up Window-based Detection
Limitations of Window-based Detection
Detection with Part-based Models
Combination Classifiers
Voting and the Generalized Hough Transform
RANSAC
Generalized Distance Transform
Learning Generic Object Category Models
Data Annotation
Learning Window-based Models
Specialized Similarity Measures and Kernels
Learning Part-based Models
Learning in the Constellation Model
Learning in the Implicit Shape Model
Learning in the Pictorial Structure Model
Example Systems: Generic Object Recognition
The Viola-Jones Face Detector
Training Process
Recognition Process
Discussion
The HOG Person Detector
Bag-of-Words Image Classification
Training Process
Recognition Process
Discussion
The Implicit Shape Model
Training Process
Recognition Process
Vote Backprojection and Top-Down Segmentation
Hypothesis Verification
Discussion
Deformable Part-based Models
Training Process
Recognition Process
Discussion
Other Considerations and Current Challenges
Benchmarks and Datasets
Context-based Recognition
Multi-Viewpoint and Multi-Aspect Recognition
Role of Video
Integrated Segmentation and Recognition
Supervision Considerations in Object Category Learning
Using Weakly Labeled Image Data
Maximizing the Use of Manual Annotations
Unsupervised Object Discovery
Language, Text, and Images
Conclusions
Bibliography
Authors' Biographies