Performance Evaluation in Content-based Image Retrieval: The Benchathlon Network
暂无分享,去创建一个
In this article, we give an overview of the main tasks involved in designing a platform for the evaluation of content-based image retrieval systems. A number of issues should be addressed from the construction of an image collection to the definition of standard performance measures. The aim here is to introduce the Benchathlon network that has been created with the aim of setting up such a platform. Sub tasks engendered by this context are listed and the organization of our collaboration detailed. While participants are allocated to specific subtasks, everyone may submit a solution to another parallel task. The last part of this article gives a list of the contribution that our Viper team has submitted to the Benchathlon network. Perspectives for these benchmarking efforts are finally given.
[1] Syed M Rahman. Design and Management of Multimedia Information Systems: Opportunities and Challenges , 2000 .
[2] Thierry Pun,et al. Performance evaluation in content-based image retrieval: overview and proposals , 2001, Pattern Recognit. Lett..
[3] Neil J. Gunther,et al. Benchmark for image retrieval using distributed systems over the Iinternet: BIRDS-I , 2000, IS&T/SPIE Electronic Imaging.
[4] Thomas Pfund,et al. Dynamic multimedia annotation tool , 2001, IS&T/SPIE Electronic Imaging.