TRAED: speech audio editing using imperfect transcripts.
Masood MasoodianBill RogersDavid WareSam McKoyPublished in: MMM (2006)
Keyphrases
- broadcast news
- spoken documents
- automatic speech recognition
- audio stream
- spoken document retrieval
- speaker identification
- spontaneous speech
- audio visual
- speech recognizer
- speech transcripts
- emotion recognition
- spoken term detection
- speech signal
- audio signals
- text to speech
- speech recognition
- audio features
- video search
- acoustic features
- speech processing
- video retrieval
- cepstral features
- hidden markov models
- automatic transcription
- prosodic features
- audio recordings
- speaker diarization
- linear predictive coding
- speech music discrimination
- digital audio
- speech synthesis
- out of vocabulary
- image editing
- acoustic signals
- human machine interaction
- spoken language
- music information retrieval
- content analysis
- multimedia data
- feature extraction
- information retrieval