Presentation 2014-12-12
A proposal for data selection in self-training based cross dataset action recognition
Takafumi SUZUKI, Yu WANG, Jien KATO, Kenji MASE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In action recognition, in order to obtain high performance classifiers, it is necessary to feed the training algorithm enough labeled data. Since labeling is a very expensive task, it is necessary to develop approaches which can efficiently reuse labeled data. In this work, we consider the task of utilizing labeled data from one dataset (source dataset) to train action classifiers for data from another completely unlabeled dataset (target dataset). We propose a novel approach for such a task by extending the well-known self-training algorithm to including two different data selecting criterions that are inspired by Support Vector Machine and k-Nearest Neighbor. The superior of our approach has been confirmed by benchmark dataset.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Action Recognition / Pattern Recognition / Self Training
Paper # PRMU2014-80
Date of Issue

Conference Information
Committee PRMU
Conference Date 2014/12/4(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 Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A proposal for data selection in self-training based cross dataset action recognition
Sub Title (in English)
Keyword(1) Action Recognition
Keyword(2) Pattern Recognition
Keyword(3) Self Training
1st Author's Name Takafumi SUZUKI
1st Author's Affiliation Nagoya University()
2nd Author's Name Yu WANG
2nd Author's Affiliation Nagoya University
3rd Author's Name Jien KATO
3rd Author's Affiliation Nagoya University
4th Author's Name Kenji MASE
4th Author's Affiliation Nagoya University
Date 2014-12-12
Paper # PRMU2014-80
Volume (vol) vol.114
Number (no) 356
Page pp.pp.-
#Pages 5
Date of Issue