Incorporate Visual Analytics to Design a Human-Centered Computing Framework for Personalized Classifier Training and Image Retrieval

Abstract

Human has always been a part of the computational loop. The goal of human-centered multimedia computing is to explicitly address human factors at all levels of multimedia computations. In this chapter, we have incorporated a novel visual analytics framework to design a human-centered multimedia computing environment. In the loop of image classifier training, our visual analytics framework can allow users to obtain better understanding of the hypotheses, thus they can further incorporate their personal preferences to make more suitable hypotheses for achieving personalized classifier training. In the loop of image retrieval, our visual analytics framework can also allow users to gain a deep insights of large-scale image collections at the first glance, so that they can specify their queries more precisely and obtain the most relevant images quickly. By supporting interactive image exploration, users can express their query intentions explicitly and our system can recommend more relevant images adaptively.