Investigation on new Mel frequency cepstral coefficients features and hyper-parameters tuning technique for bee sound recognition.
Thi-Thu-Hong PhanDong Nguyen-DoanDu Nguyen-HuuHanh Nguyen-VanThai Pham-HongPublished in: Soft Comput. (2023)
Keyphrases
- mel frequency cepstral coefficients
- acoustic features
- extracting features
- speaker identification
- feature vectors
- feature extraction
- speaker recognition
- feature set
- hyperparameters
- speech signal
- gaussian mixture model
- cross validation
- object recognition
- principal component analysis
- audio features
- incremental learning
- low level
- model selection
- image processing
- feature space
- pairwise
- automatic speech recognition
- random sampling
- parameter settings
- speech recognition
- classification accuracy
- bayesian framework
- learning algorithm
- visual features
- support vector
- speaker diarization
- dimensionality reduction