Audio-Based Video Content Analysis on User-Generated Content

Benjamin Elizalde

ICSI

Tuesday, October 8, 2013
12:30 p.m., Conference Room 5A

Abstract:

Audio-based video event detection on user-generated content (UGC) aims to find videos that show an observable event, such as a wedding ceremony or a birthday party. In a lower tier, audio concept detection aims to find a sound or concept, such as music, clapping or a cat's meow. Different events are described by different sounds. The difficulty of video content analysis on UGC lies in the lack of structure and acoustic variability of the data. The video content analysis has been explored mainly by computer vision, but it requires audio to complement the search of cues on this multimedia challenge. The talk is about our audio-based video event detection and audio concept classification systems and research. Additionally, ICSI's participation on the TRECVID MED evaluations and the latest results.

Bio:

Benjamin Elizalde joined the multimedia research group led by Dr. Gerald Friedland in 2012, on the Aladdin Project of "Multimedia Event Detection on User-Generated Content". He graduated with his masters degree at ITESM in Monterrey, Mexico. The thesis topic related to the Aladdin Project, began during his research stay at Carnegie Mellon and finished at ICSI. He also has a Bachelor in Electronic and Computer Engineering from ITESM and had the chance to work as an IT Consultant at WSS (Spain) on software projects for companies like Royal Bank Canada - Dexia and John Deere.