Audio thumbnails for spoken content without transcription based on a maximum motif coverage criterion.
Guillaume GravierNathan Souviraà-LabastieSébastien CampionFrédéric BimbotPublished in: INTERSPEECH (2014)
Keyphrases
- automatic transcription
- multimedia
- speech recognition technology
- cross modal
- speech recognition
- audio visual
- motif discovery
- image browsing
- optimization criterion
- dna sequences
- music information retrieval
- audio video
- visual data
- music score
- handwriting recognition
- signal processing
- audio features
- broadcast news
- original images
- feature selection