The Computational Visual Media Conference series is intended to provide a major international forum for exchanging novel research ideas and significant practical results both underpinning and applying Visual Media. The primary rationale for this new conference series is to target cross disciplinary research which amalgamates aspects of computer graphics, computer vision, machine learning, image processing, video processing, visualization and geometric computing. Original research is sought in areas concerning the classification, composition, retrieval, synthesis, and understanding of visual media. The previous conferences were held in Tsinghua University in 2012 and Zhejiang University in 2013.