Comparing MFCC and MPEG-7 audio features for feature extraction, maximum likelihood HMM and entropic prior HMM for sports audio classification.
Ziyou XiongRegunathan RadhakrishnanAjay DivakaranThomas S. HuangPublished in: ICASSP (5) (2003)
Keyphrases
- mel frequency cepstral coefficients
- audio features
- speech signal
- feature extraction
- hidden markov models
- speech recognition
- feature set
- speaker identification
- maximum likelihood
- acoustic features
- audio visual
- genre classification
- music genre classification
- feature vectors
- speaker recognition
- audio stream
- image classification
- automatic speech recognition
- low level
- pattern recognition
- visual features
- gaussian mixture model
- feature selection
- music information retrieval
- audio signal
- spectral features
- extracted features
- automatic music genre classification
- multimedia
- feature space
- classification accuracy
- principal component analysis
- image processing
- noisy environments
- multi modal
- expectation maximization
- speaker diarization
- audio signals
- collaborative filtering
- machine learning
- video sequences
- information extraction
- visual information
- language model
- extracting features
- em algorithm
- non stationary