File:FrameErrorsVsSegmentFramesRobustPoq.png
From SpeechWiki
Size of this preview: 800 × 431 pixels
Full resolution (1,178 × 634 pixels, file size: 17 KB, MIME type: image/png)
Stats on 463286 frames. Solid lines are on amoeba-algorithm optimized weights, dashed lines are on pre-amoeba filter weights. maximum p(q|o) such that it's p(o,q) is higher than the geometric average over the p(o,q)'s of the segment is chosen as the representative frame for the segment.
File history
Click on a date/time to view the file as it appeared at that time.
| Date/Time | Thumbnail | Dimensions | User | Comment | |
|---|---|---|---|---|---|
| current | 05:42, 14 April 2010 | | 1,178×634 (17 KB) | Arthur (Talk | contribs) | (Stats on 463286 frames. Solid lines are on amoeba-algorithm optimized weights, dashed lines are on pre-amoeba filter weights.) |
- Edit this file using an external application (See the setup instructions for more information)
File links
The following page links to this file:
