Intermix: An Interference-Based Data Augmentation and Regularization Technique for Automatic Deep Sound Classification.
Ramit SawhneyAtula Tejaswi NeerkajePublished in: ICASSP (2022)
Keyphrases
- data sets
- database
- pattern recognition
- data analysis
- data collection
- data sources
- machine learning
- high quality
- data structure
- image data
- synthetic data
- data points
- data quality
- computer systems
- support vector machine svm
- probability distribution
- data reduction
- training data
- feature extraction
- high dimensional data
- audio signal
- training samples
- unsupervised learning
- multi class
- end users
- prior knowledge
- preprocessing
- support vector
- clustering algorithm