Decision Fusion Based on Voting Technology for Content-Based Natural Image Retrieval
Tianwen Zhang · Journal of Guangxi Normal University · 2007
In content-based image retrieval,multi-feature of image can effectively make up the limitation from representation of single feature.The visual information represented by different features is different,thus, fusion of multi-feature contributes to the performance of image retrieval.In this article,image retrieval is considered as classification,and Borda count method is modified according to the attribute of image retrieval.The decisions of different classifiers are combined by sorting and ranking.The result of experiment shows that this method is practicable.