Two-class signal segmentation for speech/music detection in audio tracks.
Mouhamadou SeckFrédéric BimbotDidier ZugajBernard DelyonPublished in: EUROSPEECH (1999)
Keyphrases
- audio signals
- audio signal
- audio features
- speech music discrimination
- low signal to noise ratio
- audio recordings
- audio visual
- signal processing
- digital audio
- signal detection
- speaker identification
- music information retrieval
- broadcast news
- music genre classification
- level set
- noisy environments
- image segmentation
- automatic speech recognition
- multimedia
- voice activity detection
- speech signal
- emotion recognition
- music score
- music scores
- matched filter
- segmentation algorithm
- object detection
- audio stream
- acoustic features
- visual features
- music retrieval
- speech segments
- object segmentation
- bounding box
- medical images
- acoustic signals
- multiscale
- frequency domain
- visual information
- speech quality
- acoustic signal
- edge detection
- music collections
- speech recognition
- sound source
- content based music retrieval