Damian Borth is visiting ICSI’s Audio and Multimedia team on a postdoctoral fellowship within ICSI's German visiting agreement, which is funded through the FITweltweit program, sponsored by the BMBF and administered by the DAAD, the German Academic Exchange Program.

He first visited California during a student exchange stay at UC Santa Barbara while working toward his degree in engineering, which he received in Germany in 2003. After working in Taiwan for a couple of years, he returned to Germany to study computer science. He then received a master’s degree in 2010 and was also a visiting scholar at Columbia University in New York before receiving his PhD from the University of Kaiserlautern, Germany. While working toward his PhD, he also worked at the German Research Center for Artificial Intelligence (DFKI) under Professor Thomas Breuel. His thesis work and his research for the DFKI were on visual learning using data from the Web, specifically video from social media sites like YouTube and Flickr. In his visual learning research, he worked toward bridging the semantic gap as well as trending topic detection and visual sentiment analysis. This is useful for content-based retrieval: being able to search a database not just for video titles or tags but also for video content.

(Damian points out that his brother, who works at the visual effects company that, among other things, worked on “Star Trek Into Darkness,” creates video content, whereas Damian analyzes video content – so it runs in the family.)

Sentiment analysis, a related area of research, has in the past been applied to text and attempts to find the sentiment, or emotional meaning, of the message. But as social networks more and more frequently incorporate different media – for example, pictures embedded in Twitter messages – it’s important to understand text in relation to images, and to understand the sentiment of the image itself. Damian points to the example of the re-election image of President Obama posted to Twitter with the tweet, “Four more years.” The text in itself doesn’t convey the sentiment of the tweet; by incorporating visual analysis, it becomes clear that the tweet is conveying a positive sentiment.

Damian says he would like to continue working on multimedia opinion mining, which he’s worked on at Columbia University (where he performed sentiment analysis) and DFKI (where he studied global distributions of trending topics).

Damian enjoys playing and watching fussball (soccer to us) and was confused by the American term foosball (table soccer).