Multimedia Analysis with Deep Learning

摘要

Recently, deep learning method has been attracting more and more researchers due to its great success in various computer vision tasks. Particularly, some researchers focus on the study of multimedia analysis by deep learning method, and the research tasks mainly include the following six aspects: classification, retrieval, segmentation, tracking, detection and recommendation. As far as we know, there is not any literature conducting on survey of these studies, and it is of great significance for the community to review this subject. In this paper, we discuss the application of deep learning method in the six multimedia analysis tasks, and also point out the future directions of deep learning in multimedia analysis.