期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2013
卷号：4
期号：8-5
出版社：Seventh Sense Research Group
摘要：Active learning is a machine learning technique which chooses the most informative models for labelling and uses them as training data. It has been extensively explored in multimedia research area for reducing human annotation effort. In this article, efforts of active learning in multimedia annotation and retrieval have been surveyed .The application domains such as image or video annotation, Relevance feedback and contentbased image retrieval are mainly focussed. The principle of active learning has been briefly discussed and then sample selection criteria were analyzed. Classification models used in active learningbased multimedia annotation and retrieval, including semisupervised learning, Support vector machine has been discussed. In particular, largescale interactive multimedia annotation and analysis of human annotation with several strategies were briefly discussed.
