Social multimedia indicates the multimedia data generated and consumed under social media circumstances. As the hybrid of multimedia and social media, social multimedia enjoys advantages of both direct rich sensory simulation and efficient information access and propagation, thus having great potentials in analysis and utilization. Social multimedia computing, a multidisciplinary research and application field, has been developed to understand social multimedia content and connect the social multimedia content with users by exploiting the various social interactions. Social multimedia computing is very different from traditional and web multimedia computing. In traditional multimedia computing, the analysis focus is the multimedia content, and the goal is content understanding and application, e.g., media content analysis, semantic classification and annotation, structured media authoring. Web multimedia computing is heavily related to the WEB1.0 environment, which is dominated by broadcast media developed by professional designers for passive users. Impacted by the participatory WEB 2.0, social multimedia computing features in two perspectives: (1) Since users actively participate in the data generation as well as consumption processes, it is important to exploit the users' interaction and embed social knowledge into the loop of multimedia content understanding. (2) User is the ultimate target of social multimedia information service. Understanding the customized and personalized demands is critical to most social multimedia computing problems.