A Discriminative and Compact Audio Representation for Event Detection

Liping Jing, Bo Liu, Jaeyoung Choi, Adam Janin, Julia Bernd, Michael W. Mahoney, Gerald Friedland

Abstract

This paper presents a novel two-phase method for audio representation: Discriminative and Compact Audio Representation (DCAR). In the first phase, each audio track is modeled using a Gaussian mixture model (GMM) that includes several components to capture the variability within that track. The second phase takes into account both global structure and local structure. In this phase, the components are rendered more discriminative and compact by formulating an optimization problem on Grassmannian manifolds, which we found represents the structure of audio effectively. Experimental results on the YLI-MED dataset show that the proposed DCAR representation consistently outperforms state-of-the-art audio representations: i-vector, mv-vector, and GMM.
BibTeX
@inproceedings{jing2016discriminative,
  author = {Liping Jing and Bo Liu and Jaeyoung Choi and Adam Janin and Julia Bernd and Michael W. Mahoney and Gerald Friedland},
  title = {{A Discriminative and Compact Audio Representation for Event Detection}},
  booktitle = {ACM Multimedia},
  year = {2016},
  doi = {10.1145/2964284.2970377},
}

Loading related content...

Figure