Diversifying Plant Image Retrieval with Dimensionality Reduction

In recent years, content-based image retrieval achieved continuous development, but in the previous studies, only the relevance is cared in retrieval system, so many duplicate or near duplicate documents retrieved in response to a query and cannot satisfy the users. To solve this problem, we propose the Content-based Diversifying Plant Image Retrieval in this paper. In order to make the retrieval results have relevance and diversity, we extract plant image feature and use the relevance feedback technique based of SVM and the AP clustering algorithm. To accelerate retrieval, we use maximum variance projection (MVP) for dimensionality reduction. Experimental results show that our approach can achieve good performance.