Image Information Mining: an Accelerated Bayesian Algorithm for Data Fusion of SAR Big Data

Abstract

The paper presents a knowledge-driven content-based information mining Bayesian algorithm for data fusion of SAR Big Data. The method combines, at pixel level, the unsupervised clustering results of different number of features with a user given semantic concept. The combination has as goal to calculate the posterior probability that allows the final search. The proposed interactive learning method is able to learn different user semantic labels, which can be used to retrieve related images with a few user interactions, greatly optimizing the computational costs and over performing existing similar systems in various orders of magnitude.