Copyright Tag Retrieval of Estate Images Based on Improved SIFT
Jun Jie Yang, Shi-jiao Zhu · International Conference on Multimedia Information Networking and Security · 2013
Accompany with network widely used, a large number of online images retrieval are required. Online retrieval of house estates' copyright tags, one of applications of online image retrieval, often relies on the image's descriptors and hardly in using pixels model. In this paper, firstly, we use local descriptors to present an image and illustrate its advantages, then, a filtration system for web house estate images based on improved scale invariant feature transform (SIFT) analysis is proposed, it uses local information of image to make features from original image, and then comparison is performed to find similar images in feature space. Experiments results performed on a real web house estate image database demonstrate that the method can greatly decrease manual physical find task in estate image retrieval application.