No abstract

Cardiovascular diseases have a high morbidity, and remain the leading cause of mortality. In the past two decades, developing an intelligent auscultation system has attracted tremendous efforts from the field of signal processing and machine learning. We propose a novel framework based on wavelet representations and deep recurrent neural networks for recognising three heart sounds, i. e., normal, mild, and severe. The Heart Sounds Shenzhen corpus (n = 170) is used to validate the proposed method. The experimental results demonstrate the efficacy of the proposed method in a rigorous subject independent scenario, which can reach an unweighted average recall at 43.0 % (chance level: 33.3 %).

In this study, we propose a methodology for separating a singing voice from musical accompaniment in a monaural musical mixture. The proposed method uses robust principal component analysis (RPCA), followed by postprocessing, including median filter, morphology, and high-pass filter, to decompose the mixture. Subsequently, a deep recurrent neural network comprising two jointly optimized parallel-stacked recurrent neural networks (sRNNs) with mask layers and trained on limited data and computation is applied to the decomposed components to optimize the final estimated separated singing voice and background music to further correct misclassified or residual singing and background music in the initial separation. The experimental results of MIR-1K, ccMixter, and MUSDB18 datasets and the comparison with ten existing techniques indicate that the proposed method achieves competitive performance in monaural singing voice separation. On MUSDB18, the proposed method reaches the comparable separation quality in less training data and lower computational cost compared to the other state-of-the-art technique.

In this paper, the multiplicative syllable duration model proposed previously for Mandarin speech is extended in some aspects. First, the three basic Tone 3 patt erns (i.e., full lone, half lone and sandhi tone) are properly considered via using three different companding factors (CFs) to separate their affections. Second, the CPs of the model are analyzed in detail. Third, the syllable duration modeling method is applied 10 an automatically-segmented, SOO-speaker, telephone-speech database. Fourth, a comparative study 10 paraUelly construct an add itive syllable duration model is done.