Gong / Xu Machine Learning for Multimedia Content Analysis

1. Auflage 2007
ISBN: 978-0-387-69942-4
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, Band 30, 277 Seiten

Reihe: Multimedia Systems and Applications

This volume introduces machine learning techniques that are particularly powerful and effective for modeling multimedia data and common tasks of multimedia content analysis. It systematically covers key machine learning techniques in an intuitive fashion and demonstrates their applications through case studies. Coverage includes examples of unsupervised learning, generative models and discriminative models. In addition, the book examines Maximum Margin Markov (M3) networks, which strive to combine the advantages of both the graphical models and Support Vector Machines (SVM).

Autoren/Hrsg.

Gong, Yihong

Xu, Wei

Inhaltsverzeichnis

1;Preface;6
2;Contents;11
3;1 Introduction;16
3.1;1.1 Basic Statistical Learning Problems;17
3.2;1.2 Categorizations of Machine Learning Techniques;19
3.3;1.3 Multimedia Content Analysis;23
4;Part I Unsupervised Learning;27
4.1;2 Dimension Reduction;28
4.1.1;2.1 Objectives;28
4.1.2;2.2 Singular Value Decomposition;29
4.1.3;2.3 Independent Component Analysis;33
4.1.4;2.4 Dimension Reduction by Locally Linear Embedding;39
4.1.5;2.5 Case Study;43
4.1.6;Problems;47
4.2;3 Data Clustering Techniques;49
4.2.1;3.1 Introduction;49
4.2.2;3.2 Spectral Clustering;51
4.2.3;3.3 Data Clustering by Non-Negative Matrix Factorization;63
4.2.4;3.4 Spectral vs. NMF;71
4.2.5;3.5 Case Study: Document Clustering Using Spectral and NMF Clustering Techniques;73
4.2.6;Problems;80
5;Part II Generative Graphical Models;83
5.1;4 Introduction of Graphical Models;84
5.1.1;4.1 Directed Graphical Model;85
5.1.2;4.2 Undirected Graphical Model;88
5.1.3;4.3 Generative vs. Discriminative;90
5.1.4;4.4 Content of Part II;91
5.2;5 Markov Chains and Monte Carlo Simulation;92
5.2.1;5.1 Discrete-Time Markov Chain;92
5.2.2;5.2 Canonical Representation;95
5.2.3;5.3 Definitions and Terminologies;99
5.2.4;5.4 Stationary Distribution;102
5.2.5;5.5 Long Run Behavior and Convergence Rate;105
5.2.6;5.6 Markov Chain Monte Carlo Simulation;111
5.2.7;Problems;123
5.3;6 Markov Random Fields and Gibbs Sampling;126
5.3.1;6.1 Markov Random Fields;126
5.3.2;6.2 Gibbs Distributions;128
5.3.3;6.3 Gibbs – Markov Equivalence;131
5.3.4;6.4 Gibbs Sampling;134
5.3.5;6.5 Simulated Annealing;137
5.3.6;6.6 Case Study: Video Foreground Object Segmentation by MRF;144
5.3.7;Problems;157
5.4;7 Hidden Markov Models;159
5.4.1;7.1 Markov Chains vs. Hidden Markov Models;159
5.4.2;7.2 Three Basic Problems for HMMs;163
5.4.3;7.3 Solution to Likelihood Computation;164
5.4.4;7.4 Solution to Finding Likeliest State Sequence;168
5.4.5;7.5 Solution to HMM Training;170
5.4.6;7.6 Expectation-Maximization Algorithm and its Variances;172
5.4.7;7.7 Case Study: Baseball Highlight Detection Using HMMs;177
5.4.8;Problems;185
5.5;8 Inference and Learning for General Graphical Models;188
5.5.1;8.1 Introduction;188
5.5.2;8.2 Sum-product algorithm;191
5.5.3;8.3 Max-product algorithm;197
5.5.4;8.4 Approximate inference;198
5.5.5;8.5 Learning;200
5.5.6;Problems;205
6;Part III Discriminative Graphical Models;207
6.1;9 Maximum Entropy Model and Conditional Random Field;208
6.1.1;9.1 Overview of Maximum Entropy Model;209
6.1.2;9.2 Maximum Entropy Framework;211
6.1.3;9.3 Comparison to Generative Models;217
6.1.4;9.4 Relation to Conditional Random Field;220
6.1.5;9.5 Feature Selection;222
6.1.6;9.6 Case Study: Baseball Highlight Detection Using Maximum Entropy Model;224
6.1.7;Problems;239
6.2;10 Max-Margin Classifications;241
6.2.1;10.1 Support Vector Machines (SVMs);242
6.2.2;9);257
6.2.3;9,;257
6.2.4;10.2 Maximum Margin Markov Networks;263
6.2.5;Problems;270
7;A Appendix;273
8;References;274
9;Index;280