Propelled by the rapid advancement of massive storage devices and delivery technologies, both the amount and the diversity of consumer-produced multimedia content on the web are growing at an incredible rate. Multimedia sources such as image, audio and video are almost unstructured or semi-structured data by nature, which leads to poor accessibility without using efficient and effective technologies and solutions. Recent progress on data mining and visualization techniques stimulates researchers to seek methods for large-scale multimedia content retrieval and analysis. As a relatively new field of study, visual analytics brings together new computational and theory-based tools to generate insightful results and findings from massive and complex data. By coupling interactive visual representations with underlying analytical processes, this science of analytical reasoning is becoming one of the fastest ways for people to explore and understand large multimedia collections. This special issue will target the most advanced multimedia mining topics and advocates a promising combination of automatic analysis and visualization techniques to deal with the challenges of multimedia processing in large-scale environments.