Incorporate support vector machines to content-based image retrieval with relevance feedback

摘要: By using relevance feedback, content-based image retrieval (CBIR) allows the user to retrieve images interactively. Beginning with a coarse query, can select most relevant and provide weight of preference for each refine query. The high level concept borne by perception subjectivity be automatically captured system some degree. This paper proposes an approach utilize both positive negative feedbacks retrieval. Support vector machines (SVM) is applied classifying images. SVM learning results are used update weights releases from manually providing example. Experimental show that proposed has improvement over previous (Rui et al. 1997) uses examples only.