Integrating Graphics and Vision for Object Recognition

by Stevens, Mark R.; Beveridge, J. Ross

• ISBN13: 9780792372073
• ISBN10: 0792372077
• Format: Hardcover
• Copyright: 2000-10-01
• Publisher: Kluwer Academic Pub

Summary

Integrating Graphics and Vision for Object Recognition serves as a reference for electrical engineers and computer scientists researching computer vision or computer graphics. Computer graphics and computer vision can be viewed as different sides of the same coin. In graphics, algorithms are given knowledge about the world in the form of models, cameras, lighting, etc., and infer (or render) an image of a scene. In vision, the process is the exact opposite: algorithms are presented with an image, and infer (or interpret) the configuration of the world. This work focuses on using computer graphics to interpret camera images: using iterative rendering to predict what should be visible by the camera and then testing and refining that hypothesis. Features of the book include: Many illustrations to supplement the text; A novel approach to the integration of graphics and vision; Genetic algorithms for vision; Innovations in closed loop object recognition. Integrating Graphics and Vision for Object Recognition will be of interest to research scientists and practitioners working in fields related to the topic. It may also be used as an advanced-level graduate text.

Table of Contents

List of Figures	vii
List of Tables	xi
Introduction	1	(10)
What Does it mean to Interpret a Scene?	2	(2)
Using Partial Scene Models for Object Recognition	4	(3)
Example Problem Domains	7	(1)
Assumptions Made	8	(1)
Overview	9	(2)
Previous Work	11	(22)
Contributing Individuals	13	(2)
Our Place in the Pipeline	15	(1)
Geometric Feature Matching	15	(6)
Augmented Geometric Matching	21	(4)
Appearance Matching	25	(4)
RMR: Fusing Geometry and Appearance	29	(4)
Render: Predicting Scenes	33	(24)
Overview of the RMR Dataset	35	(2)
Formally Defining the Scene Configuration	37	(7)
The Model Representation	44	(4)
Rendering Objects Using the Scene Configuration	48	(7)
Predicting a Simple Background Model	55	(2)
Match: Comparing Images	57	(40)
Defining Specific Error Functions	59	(21)
Comparing Error Functions	80	(17)
Refine: Iterative Search	97	(34)
Desirable Properties of a Search Algorithm	98	(3)
Defining Specific Search Algorithms	101	(9)
Comparing Search Algorithms	110	(21)
Evaluation	131	(22)
An Overview of the Results	132	(1)
Evaluating Performance	133	(12)
Relating Dependent and Independent Variables	145	(5)
Summary	150	(3)
Conclusions	153	(4)
Future Work	155	(2)
Appendices	157	(22)
A-- Generating Scene Hypotheses	157	(22)
1. Object Detection and Pose Indexing	157	(2)
2. Detection based on Color Decision Trees	159	(4)
3. Pose Indexing	163	(16)
Index	179