摘要

Following the development of the internet, there is more and more image database built for supporting indexing, searching and browsing. This work designs a retrieval scheme which has a high efficiency for the image with complex background. We utility the object model to train the Support Vector Machine classifier. In the processing of retrieval, firstly it need segment doubtful object from the whole image, then use the classifier to recognize. The results of exercises illustrate this scheme has quicker speed in recognizing than use the traditional scheme, especially for the retrieval accuracy for the image with complex background.

  • 出版日期2012

全文