The explosive increase of multimedia data (i.e., text, image, and video) on the Internet has brought the great challenge of how to effectively index, retrieve and organize these resources. Much research attention has been paid to discover semantic concepts in multimedia data. This special issue aims to collect recent state-of-the-art achievement on semantic concept discovery, especially the work devoted to several new challenges in this field. Another focus of this special issue will be recent advances on semantic event detection, which refers to detect unseen event without any training data. To achieve this goal, a bunch of concept classifiers should be trained beforehand. How to efficiently train these classifiers with millions of videos is still an open problem.

This special issue targets the researchers and practitioners from both academia and industry. In total, 81 submissions have been received. After three rounds of reviews, we finally accepted 40 of them. There are around 40 reviewers involved in the review process.