Exploiting generalized discriminative multiple instance learning for multimedia semantic concept detection

Sheng Gao

doi:10.1016/j.patcog.2008.03.029

Return

Exploiting generalized discriminative multiple instance learning for multimedia semantic concept detection

2008-10-01

PRE

高胜 (Sheng Gao) *

孙

孙启彬 (Qibin Sun)

DOI：10.1016/j.patcog.2008.03.029

Original

Original request for help

Save

Abstract

En 中文

A generalized discriminative multiple instance learning (GDMIL) algorithm is presented to train the classifier in the condition of vague annotation of training samples GDMIL not only inherits the original MIL's capability of automatically weighting the instances in the bag according to their relevance to the concept but also integrates generative models using discriminative training. It is evaluated on the task of multimedia semantic concept detection using the development data set of TRECVID 2005. The experimental results show GDMIL outperforms the baseline systems trained on MIL with diverse density and expectation-maximization diverse density and the system without MIL. (C) 2008 Elsevier Ltd. All rights reserved.

Keywords:

multiple instance learning

discriminative training

semantic concept detection

area under the ROC curve

classification accuracy

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.