This paper presents and explores a robust deep learning framework for auscultation analysis. This aims to classify anomalies in respiratory cycles and detect diseases, from respiratory sound recordings. The framework begins with front-end feature extraction that transforms input sound into a spectrogram representation. Then, a back-end deep learning network is used to classify the spectrogram features into categories of respiratory anomaly cycles or diseases. Experiments, conducted over the ICBHI benchmark dataset of respiratory sounds, confirm three main contributions towards respiratory- sound analysis. Firstly, we carry out an extensive exploration of the effect of spectrogram types, spectral-time resolution, overlapping/non-overlapping windows, and data augmentation on final prediction accuracy. This leads us to propose a novel deep learning system, built on the proposed framework, which outperforms current state-of-the-art methods. Finally, we apply a Teacher-Student scheme to achieve a trade-off between model performance and model complexity which holds promise for building real-time applications.

Generative adversarial networks (GAN) have recently been shown to be efficient for speech enhancement. However, most, if not all, existing speech enhancement GANs (SEGAN) make use of a single generator to perform one-stage enhancement mapping. In this work, we propose to use multiple generators that are chained to perform multi-stage enhancement mapping, which gradually refines the noisy input signals in a stage-wise fashion. Furthermore, we study two scenarios: (1) the generators share their parameters and (2) the generators' parameters are independent. The former constrains the generators to learn a common mapping that is iteratively applied at all enhancement stages and results in a small model footprint. On the contrary, the latter allows the generators to flexibly learn different enhancement mappings at different stages of the network at the cost of an increased model size. We demonstrate that the proposed multi-stage enhancement approach outperforms the one-stage SEGAN baseline, where the independent generators lead to more favorable results than the tied generators. The source code is available at http://github.com/pquochuy/idsegan.

Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when developing sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases.

Objective: Brain waves vary between people. This work aims to improve automatic sleep staging for longitudinal sleep monitoring via personalization of algorithms based on individual characteristics extracted from sleep data recorded during the first night. Approach: As data from a single night are very small, thereby making model training difficult, we propose a Kullback–Leibler (KL) divergence regularized transfer learning approach to address this problem. We employ the pretrained SeqSleepNet (i.e. the subject independent model) as a starting point and finetune it with the single-night personalization data to derive the personalized model. This is done by adding the KL divergence between the output of the subject independent model and it of the personalized model to the loss function during finetuning. In effect, KL-divergence regularization prevents the personalized model from overfitting to the single-night data and straying too far away from the subject independent model. Main results: Experimental results on the Sleep-EDF Expanded database consisting of 75 subjects show that sleep staging personalization with single-night data is possible with help of the proposed KL-divergence regularization. On average, we achieve a personalized sleep staging accuracy of 79.6%, a Cohen’s kappa of 0.706, a macro F1-score of 73.0%, a sensitivity of 71.8%, and a specificity of 94.2%. Significance: We find both that the approach is robust against overfitting and that it improves the accuracy by 4.5 percentage points compared to the baseline method without personalization and 2.2 percentage points compared to it with personalization but without regularization.

The human auditory system is very well matched to both human speech and environmental sounds. Therefore, the question arises whether human speech material may provide useful information for training systems for analyzing nonspeech audio signals, e.g., in a classification task. In order to answer this question, we consider speech patterns as basic acoustic concepts, which embody and represent the target nonspeech signal. To find out how similar the nonspeech signal is to speech, we classify it with a classifier trained on the speech patterns and use the classification posteriors to represent the closeness to the speech bases. The speech similarities are finally employed as a descriptor to represent the target signal. We further show that a better descriptor can be obtained by learning to organize the speech categories hierarchically with a tree structure. Furthermore, these descriptors are generic. That is, once the speech classifier has been learned, it can be employed as a feature extractor for different datasets without retraining. Lastly, we propose an algorithm to select a sufficient subset, which provides an approximate representation capability of the entire set of available speech patterns. We conduct experiments for the application of audio event analysis. Phone triplets from the TIMIT dataset were used as speech patterns to learn the descriptors for audio events of three different datasets with different complexity, including UPC-TALP, Freiburg-106, and NAR. The experimental results on the event classification task show that a good performance can be easily obtained even if a simple linear classifier is used. Furthermore, fusion of the learned descriptors as an additional source leads to state-of-the-art performance on all the three target datasets.

Closed-room scenarios are characterized by reverberation, which decreases the performance of applications such as hands-free teleconferencing and multichannel sound reproduction. However, exact knowledge of the sound field inside a volume of interest enables the compensation of room effects and allows for a performance improvement within a wide range of applications. The sampling of sound fields involves the measurement of spatially dependent room impulse responses, where the Nyquist-Shannon sampling theorem applies in the temporal and spatial domains. The spatial measurement often requires a huge number of sampling points and entails other difficulties, such as the need for exact calibration of a large number of microphones. In this paper, a method for measuring sound fields using moving microphones is presented. The number of microphones is customizable, allowing for a tradeoff between hardware effort and measurement time. The goal is to reconstruct room impulse responses on a regular grid from data acquired with microphones between grid positions, in general. For this, the sound field at equidistant positions is related to the measurements taken along the microphone trajectories via spatial interpolation. The benefits of using perfect sequences for excitation, a multigrid recovery, and the prospects for reconstruction by compressed sensing are presented.