AugECE – A framework for audio signal extraction, classification, and enhancement
Classification, and Enhancement) that accomplishes the extraction, classification, and enhancement of audio signals. Specifically, AugECE first extracts the signals of interests from a mixture of audio signals via independent component analysis (ICA); AugECE then classifies them by k-nearest neighbor (kNN) to match their corresponding labels; last, wavelet-based hidden Markov models (HMM) and principal component analysis (PCA) are further employed to enhance the quality of the extracted signals by reducing background noise and/or residual signals originated from other sources as well. A high-level schematic diagram is illustrated in Fig. 1.