Presentation 2011-09-05
Feature Selection for Object Tracking based on Online Real Boosting
Takayoshi YAMASHITA, Hironobu FUJIYOSHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Recently, Boosting algorithms like AdaBoost and Real AdaBoost are used in online learning. The weak classifiers for online boosting are selected in an iterative manner in Adaboost based Online Boosting. And due to this, the computing cost increases for online training and it might also over-fit to the current sample. On the other hand, Online Real Boosting based on Real Adaboost groups the weak classifier, and selects an optimal weak classifier from each group. Besides reducing the computational cost, it prevents overfitting. The grouping process for feature selection is yet to be discussed. In this paper, we discuss the various grouping process and its effect on the tracking performance. As a result, it is observed that feature type based grouping of the weak classifiers results in the best performance.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Online Real Boosting / object tracking / feature selection / grouping / random
Paper # PRMU2011-68,IBISML2011-27
Date of Issue

Conference Information
Committee IBISML
Conference Date 2011/8/29(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Information-Based Induction Sciences and Machine Learning (IBISML)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Feature Selection for Object Tracking based on Online Real Boosting
Sub Title (in English)
Keyword(1) Online Real Boosting
Keyword(2) object tracking
Keyword(3) feature selection
Keyword(4) grouping
Keyword(5) random
1st Author's Name Takayoshi YAMASHITA
1st Author's Affiliation OMRON Corporation()
2nd Author's Name Hironobu FUJIYOSHI
2nd Author's Affiliation Chubu University
Date 2011-09-05
Paper # PRMU2011-68,IBISML2011-27
Volume (vol) vol.111
Number (no) 194
Page pp.pp.-
#Pages 7
Date of Issue