In this paper, we propose a new clustering model for speaker diarization. A major problem with using greedy agglomerative hierarchical clustering for speaker diarization is that they do not guarantee an optimal solution. We propose a new clustering model, by redefining clustering as a problem of Integer Linear Programming (ILP). Thus an ILP solver can be used which searches the solution of speaker clustering over the whole problem. The experiments were conducted on the corpus of French broadcast news ESTER-2. With this new clustering, the DER decreases by 2.43 points.
Cite as: Rouvier, M., Meignier, S. (2012) A global optimization framework for speaker diarization. Proc. The Speaker and Language Recognition Workshop (Odyssey 2012), 146-150
@inproceedings{rouvier12_odyssey,
author={Mickael Rouvier and Sylvain Meignier},
title={{A global optimization framework for speaker diarization}},
year=2012,
booktitle={Proc. The Speaker and Language Recognition Workshop (Odyssey 2012)},
pages={146--150}
}