SVM-Based Detection of Misannotated Words in Read Speech Corpora.
Jindrich MatousekDaniel TihelkaPublished in: TSD (2013)
Keyphrases
- text recognition
- text corpora
- object detection
- natural language processing
- detection method
- n gram
- training corpus
- spoken words
- speech recognition errors
- speech corpus
- continuous speech recognition
- text corpus
- spontaneous speech
- speech recognition
- support vector machine svm
- false positives
- grapheme to phoneme conversion
- noisy environments
- keywords
- word sense disambiguation
- word frequency
- spoken document retrieval
- text documents
- lexical features
- speech recognition systems
- conversational speech
- detection algorithm