Social multimedia refers to the multimedia content generated by social network users for social interactions. The increasing popularity of online social networks accumulates large amount of social network activity records, which makes the analysis of online social activities possible. The large-scale data have attracted people from both industrial and academic to mine interesting patterns from the hidden signals in the online users' activities. Researchers have successfully employed the signals extracted from social network content to finish tasks in a wide range of applications, including real world event prediction and content recommendation. From this perspective, we can view social multimedia as sensors, which provide online signals pulsing people's real word activities. One of the main features for social networks is social, where people post multimedia content intending to express their opinions and to communicate with other users.
This talk will try to analyze the current research works on social multimedia analysis, particularly on visual sentiment and emotion analysis. We investigate the new approaches to analyzing online opinions towards different topics or tasks. In our research, started from the presidential election, we have identified one of the most important factors of social multimedia including sentiment analysis. We argue that as one of the most important signals from online social networks, sentiment and emotion analysis are competitive regarding monitoring and predicting online social activities. In this talk, we summarize our results on both visual sentiment and emotion analysis as well as other related topics.