Magazine Image Retrieval with Camera-Phone

摘要

In this paper, we describe an approach for image-based retrieval with a camera-phone. We begin by getting the features in salient region of images. Then we quantify each image to a vector using the clustering-based bag-of-words model and sparse matrix algorithm. Finally, the invert document algorithm is used for speeding up the real-time query. The result of experiment shows the high efficiency in precision, query time and memory.