You might also like

More about Multimedia Data Mining and Analysis from Motion Contents

• Book details
• Synopsis
• Trending among others

You get 142 loyalty points

Book synopsis

Object motion in multimedia database contains§important information for data content analysis§especially event detection. The motion content in its§raw form is real-value multidimensional time series.§Processing and analyzing this kind of data is not§trivial since most standard machine learning§algorithms can only be applied to data in vector§space. In this work, we explore the techniques to§solve three problems related to motion cues in§multimedia databases: object motion estimation,§motion representation, and motion analysis to§recognize activities and events.§In this book we first discuss our work on object§detection and tracking, which estimate the object§motion. We then introduce our edit distance based§approach to measure the similarity between motion§trajectories in its raw form. We also present our§work on a novel trajectory representation framework,§"bag of segments," by which trajectories are§transformed to a frequency in vector space so that§most traditional machine learning algorithms can be§directly applied to the motion trajectory data.§Finally, we introduce our work on using the Granger§causality test to analyze multi-object interactions§from motion.

Book details

Book category Books in English Computing & information technology Information technology: general issues

• Full title: Multimedia Data Mining and Analysis from Motion Contents
• Author: Yue Zhou
• Language: English
• Binding: Paperback
• Number of pages: 136
• EAN: 9783639102802
• ISBN: 3639102800
• ID: 06819268
• Publisher: VDM Verlag
• Weight: 191 g
• Dimensions: 229 × 152 × 7 mm
• Date of publishing: 06. November 2008

Trending among others

---

---

---