Baseline For Track2 Using Zrtools

Juan, Benjumea; Emmanuel, Dupoux; Karadayi, Julien; Cao, Xuan-Nga; Ewan, Dunbar; Mathieu, Bernard;
  • Publisher: Zenodo
  • Related identifiers: doi: 10.5281/zenodo.815363
  • Subject: zero resource speech challenge, feature extraction, unsupervised term discovery, new paradigm | zenodo | Uncategorized

<p>For the baseline model, the system described on Jansen &amp; van Durme (2011) on PLP features was used. It performs DTW matching and uses random projections for increasing efficiency, and uses connected-component graph clustering as a second step. </p>
Share - Bookmark