Difference between revisions of "Robot Perception"

From Robot Intelligence
Jump to: navigation, search
(Human Tracking and Behavior Analysis)
(Object Recognition)
 
(16 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
=Object Recognition=
 
=Object Recognition=
* Page link to Dr. Kim
+
* [http://www.robot-intelligence.kr/index.php/Object_Recognition Object Recognition]
 +
<br />
  
 
=Human Tracking and Behavior Analysis=
 
=Human Tracking and Behavior Analysis=
===Concept===
+
==Concept==
 
* For human tracking, we are studying a method of tracking-by-detection, which use the combined visual features such as color, LBP and intensity.
 
* For human tracking, we are studying a method of tracking-by-detection, which use the combined visual features such as color, LBP and intensity.
 
* For human activity analysis, SGT(Situation Graph Tree) is introduced for easily utilizing the domain knowledge about abnormal behaviors.
 
* For human activity analysis, SGT(Situation Graph Tree) is introduced for easily utilizing the domain knowledge about abnormal behaviors.
 
* Our overall s/w schema is as follows:
 
* Our overall s/w schema is as follows:
 +
::[[File:behavior-analysis.jpg|800px|left]]
 +
<br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/><br/>
  
===Related Papers===
+
==Related Papers==
 +
*Chansu Kim and Sung-Kee Park,"Modified Particle Filtering using Foreground Separation and Confidence for Object Tracking," IEEE Advanced Video and Signal-based Surveillance, 2016.

Latest revision as of 19:26, 17 June 2016

Object Recognition


Human Tracking and Behavior Analysis

Concept

  • For human tracking, we are studying a method of tracking-by-detection, which use the combined visual features such as color, LBP and intensity.
  • For human activity analysis, SGT(Situation Graph Tree) is introduced for easily utilizing the domain knowledge about abnormal behaviors.
  • Our overall s/w schema is as follows:
Behavior-analysis.jpg






















Related Papers

  • Chansu Kim and Sung-Kee Park,"Modified Particle Filtering using Foreground Separation and Confidence for Object Tracking," IEEE Advanced Video and Signal-based Surveillance, 2016.