There is No Data Like Less Data: Percepts for Video Concept Detection on Consumer-Produced Media
Title | There is No Data Like Less Data: Percepts for Video Concept Detection on Consumer-Produced Media |
Publication Type | Technical Report |
Year of Publication | 2012 |
Authors | Friedland, G., Elizalde B. Martinez, Lei H., & Divakaran A. |
Other Numbers | 3285 |
Abstract | Video concept detection aims to find videos that show a certain event described as a high-level concept, e.g., wedding ceremony or changing a tire. This paper presents a theoretical framework and experimental evidence to suggest that video concept detection on consumer-produced videos can be performed by what we call percepts, i.e., a set of observable units with Zipfian distribution. We present an unsupervised approach to extract percepts from audio tracks, which we then use to perform experiments to provide evidence for the validity of the proposed theoretical framework using the TrecVID MED 2011 dataset. The approach suggests selecting the most relevant percepts automatically, thereby actually reducing the amount of training data. We show that our framework provides a highly usable foundation for doing video retrieval on consumer-produced content and is applicable for acoustic, visual, as well as multimodal content analysis. |
Acknowledgment | Supported by the Intelligence Advanced Research Projects Activity (IARPA) via Department ofInterior National Business Center contract number D11PC20066. The U.S. Government isauthorized to reproduce and distribute reprints for Governmental purposes notwithstandingany copyright annotation thereon. The views and conclusion contained herein are those of theauthors and should not be interpreted as necessarily representing the official policies orendorsement, either expressed or implied, of IARPA, DOI/NBC, or the U.S. Government. |
URL | http://www.icsi.berkeley.edu/pubs/techreports/TR-12-006.pdf |
Bibliographic Notes | ICSI Technical Report TR-12-006 |
Abbreviated Authors | G. Friedland, B. Martinez Elizalde, H. Lei, and A. Divakaran |
ICSI Research Group | Audio and Multimedia |
ICSI Publication Type | Technical Report |