Unsupervised training of detection threshold for polyphonic musical note tracking based on event periodicity.
Tiago Fernandes TavaresJayme Garcia Arnal BarbedoRomis AttuxAmauri LopesPublished in: ICASSP (2013)
Keyphrases
- audio recordings
- event detection
- multi target tracking
- human detection
- face detection and tracking
- supervised learning
- particle filter
- music retrieval
- music information retrieval
- detecting and tracking multiple
- real time
- crowded scenes
- threshold selection
- joint detection
- detection method
- unsupervised learning
- false detections
- musical instrument
- object detection
- video scene
- supervised training
- false positives
- training process
- person detection
- false alarms
- detection accuracy
- image sequences
- video surveillance
- appearance model
- semi supervised
- object tracking
- detection algorithm
- neural network
- training phase
- video sequences
- active learning
- text classification
- training examples
- supervised methods
- motion model
- detection rate
- visual tracking
- cluttered background
- robust tracking