Please use this identifier to cite or link to this item:
Title: Multi-graph based active learning for interactive video retrieval
Keywords: active learning, video retrieval, graph-based learning
Issue Date: 6-Jun-2009
Citation: ZHANG XIAOMING (2009-06-06). Multi-graph based active learning for interactive video retrieval. ScholarBank@NUS Repository.
Abstract: Active learning and semi-supervised learning are important machine learning techniques when labeled data is scarce or expensive to obtain. We employ a graph based semi-supervised learning method where a node in the graph represents a video shot and they are connected with edges weighted similarities. The objective is to define a smooth function that assigns a score to each node such that similar nodes have similar scores. Then we propose two fusion methods to combine multiple graphs associated with different features in order to incorporate different modalities. We apply active learning methods to select the most informative samples according to the graph structure and the current state of learning model. Experiments demonstrated the effectiveness of multi-graph based active learning method. The result on TRECVID data set shows our system achieves an MAP of 0.41 which is better than other state-of-the-arts interactive video retrieval systems.
Appears in Collections:Master's Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
thesis.pdf652.84 kBAdobe PDF



Page view(s)

checked on Nov 25, 2018


checked on Nov 25, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.