A new method for understanding sequences of utterances by multiple speakers.
Noboru MiyazakiJun-ichi HirasawaMikio NakanoKiyoaki AikawaPublished in: INTERSPEECH (2000)
Keyphrases
- significant improvement
- experimental evaluation
- preprocessing
- dynamic programming
- high accuracy
- similarity measure
- data sets
- cost function
- mutual information
- detection method
- optimization method
- fully automatic
- high precision
- synthetic data
- theoretical analysis
- markov random field
- image registration
- prior knowledge
- computational complexity
- support vector