This paper presents a novel approach based on the direct use of deep neural networks to approximate wavelet sub-bands for remote sensing (RS) image scene classification in the JPEG 2000 compressed domain. The proposed approach consists of two main steps. The first step aims to approximate the finer level wavelet sub-bands. To this end, we introduce a novel Deep Neural Network approach that utilizes the coarser level binary decoded wavelet sub-bands to approximate the finer level wavelet sub-bands (the image itself) through a series of deconvolutional layers. The second step aims to describe the high-level semantic content of the approximated wavelet sub- bands and to perform scene classification based on the learnt descriptors. This is achieved by: i) a series of convolutional layers for the extraction of descriptors which models the approximated sub-bands; and ii) fully connected layers for the RS image scene classification. Then, we introduce a loss function that allows to learn the parameters of both steps in an end-to-end trainable and unified neural network. The proposed approach requires only the coarser level wavelet sub-bands as input and thus minimizes the amount of decompression applied to the compressed RS images. Experimental results show the effectiveness of the proposed approach in terms of classification accuracy and reduced computational time when compared to the conventional use of Convolutional Neural Networks within the JPEG 2000 compressed domain.
This paper presents a novel content-based image search and retrieval (CBIR) system that achieves coarse to fine remote sensing (RS) image description and retrieval in JPEG 2000 compressed domain. The proposed system initially: i) decodes the code-streams associated to the coarse (i.e., the lowest) wavelet resolution, and ii) discards the most irrelevant images to the query image that are selected based on the similarities estimated among the coarse resolution features of the query image and those of the archive images. Then, the code-streams associated to the sub-sequent resolution of the remaining images in the archive are decoded and the most irrelevant images are selected by considering the features associated to both resolutions. This is achieved by estimating the similarities between the query image and remaining images by giving higher weights to the features associated to the finer resolution while assigning lower weights to those related to the coarse resolution. To this end, the pyramid match kernel similarity measure is exploited. These processes are iterated until the code-streams associated to the highest wavelet resolution are decoded only for a very small set of images. By this way, the proposed system exploits a multiresolution and hierarchical feature space and accomplish an adaptive RS CBIR with significantly reduced retrieval time. Experimental results obtained on an archive of aerial images confirm the effectiveness of the proposed system in terms of retrieval accuracy and time when compared to the standard CBIR systems.