Skip to main content

Model-based Visual Tracking: The OpenTL Framework

Model-based Visual Tracking: The OpenTL Framework

Giorgio Panin

ISBN: 978-0-470-87613-8

May 2011

318 pages

Select type: Hardcover

Product not available for purchase

Description

This book has two main goals: to provide a unifed and structured overview of this growing field, as well as to propose a corresponding software framework, the OpenTL library, developed by the author and his working group at TUM-Informatik.

The main objective of this work is to show, how most real-world application scenarios can be naturally cast into a common description vocabulary, and therefore implemented and tested in a fully modular and scalable way, through the defnition of a layered, object-oriented software architecture.The resulting architecture covers in a seamless way all processing levels, from raw data acquisition up to model-based object detection and sequential localization, and defines, at the application level, what we call the tracking pipeline. Within this framework, extensive use of graphics hardware (GPU computing) as well as distributed processing, allows real-time performances for complex models and sensory systems.

Preface.

1 Introduction.

1.1 Overview of the Problem.

1.2 General Tracking System Prototype.

1.3 The Tracking Pipeline.

2 Model Representation.

2.1 Camera Model.

2.2 Object Model.

2.3 Mapping Between Object and Sensor Spaces.

2.4 Object Dynamics.

3 The Visual Modality Abstraction.

3.1 Preprocessing.

3.2 Sampling and Updating Reference Features.

3.3 Model Matching with the Image Data.

3.4 Data Fusion Across Multiple Modalities and Cameras.

4 Examples of Visual Modalities.

4.1 Color Statistics.

4.2 Background Subtraction.

4.3 Blobs.

4.4 Model Contours.

4.5 Keypoints.

4.6 Motion.

4.7 Templates.

5 Recursive State-Space Estimation.

5.1 Target-State Distribution.

5.2 MLE and MAP Estimation.

5.3 Gaussian Filters.

5.4 Monte Carlo Filters.

5.5 Grid Filters.

6 Examples of Target Detectors.

6.1 Blob Clustering.

6.2 AdaBoost Classifiers.

6.3 Geometric Hashing.

6.4 Monte Carlo sampling.

6.5 Invariant Keypoints.

7 Building Applications with OpenTL.

7.1 Functional Architecture of OpenTL.

7.2 Building a Tutorial Application with OpenTL.

7.3 Other Application Examples.

Appendix A: Pose Estimation.

A.1 Point Correspondences.

A.2 Line Correspondences.

A.3 Point and Line Correspondences.

A.4 Computation of the Projective DLT Matrices.

Appendix B: Pose Representations.

B.1 Poses Without Rotation.

B.2 Parametrizing Rotations.

B.3 Poses with Rotation and Uniform Scale.

B.4 Affinity.

B.5 Poses with Rotation and Nonuniform Scale.

B.6 General Homography: The DLT Algorithm.

Nomenclature.

Bibliography.

Index.