Visual descriptors for content-based retrieval of remote sensing images

Research paper by Paolo Napoletano

Indexed on: 02 Feb '16Published on: 02 Feb '16Published in: Computer Science - Computer Vision and Pattern Recognition


In this paper we present an extensive evaluation of visual descriptors for the content-based retrieval of remote sensing images. The evaluation includes global, local, and Convolutional Neural Network (CNNs) features coupled with three different Content-Based Image Retrieval schemas. We conducted all the experiments on two publicly available datasets: the 21-class UC Merced Land Use/Land Cover data set and 19-class High-resolution Satellite Scene dataset. Results demonstrate that features extracted from CNNs are the best performing whatever is the retrieval schema adopted. Local descriptors perform better than CNN-based descriptors only when dealing with images that contain fine-grained textures or objects.