arXiv subjects

Ronald M. Aarts

Publications and source records attributed to Ronald M. Aarts.

10 recordsLinked to original sources

arXiv2026

Hybrid Real- and Complex-Valued Neural Network Architecture for Speech Enhancement

This paper investigates hybrid real- and complex-valued neural networks for monaural speech enhancement. While complex-valued models can process time-frequency representations natively, they often increase computational cost and can be inefficient in small-model regimes. We therefore study a matched-parameter hybrid architecture that combines a real-valued magnitude-mask branch with a complex-valued additive correction branch, coupled via domain conversion functions at the bottleneck. The approach is applied to convolutional denoising autoencoder and convolutional-recurrent network architectures. Averaged over four SNRs, the hybrid models improve intelligibility and quality over the considered counterparts, while substantially reducing the number of operations.

eess.AS↗

arXiv2026

Physics-Guided Variational Model for Unsupervised Sound Source Tracking

Sound source tracking is commonly performed using classical array-processing algorithms, while machine-learning approaches typically rely on precise source position labels that are expensive or impractical to obtain. This paper introduces a physics-guided variational model capable of fully unsupervised single-source sound source tracking. The method combines a variational encoder with a physics-based decoder that injects geometric constraints into the latent space through analytically derived pairwise time-delay likelihoods. Without requiring ground-truth labels, the model learns to estimate source directions directly from microphone array signals. Experiments on real-world data demonstrate that the proposed approach outperforms traditional baselines and achieves accuracy and computational complexity comparable to state-of-the-art supervised models. We further show that the method generalizes well to mismatched array geometries and exhibits strong robustness to corrupted microphone position metadata. Finally, we outline a natural extension of the approach to multi-source tracking and present the theoretical modifications required to support it.

eess.AS↗

arXiv2026

Unsupervised Variational Acoustic Clustering

We propose an unsupervised variational acoustic clustering model for clustering audio data in the time-frequency domain. The model leverages variational inference, extended to an autoencoder framework, with a Gaussian mixture model as a prior for the latent space. Specifically designed for audio applications, we introduce a convolutional-recurrent variational autoencoder optimized for efficient time-frequency processing. Our experimental results considering a spoken digits dataset demonstrate a significant improvement in accuracy and clustering performance compared to traditional methods, showcasing the model's enhanced ability to capture complex audio patterns.

eess.AS↗

arXiv2026

Categorical Unsupervised Variational Acoustic Clustering

We propose a categorical approach for unsupervised variational acoustic clustering of audio data in the time-frequency domain. The consideration of a categorical distribution enforces sharper clustering even when data points strongly overlap in time and frequency, which is the case for most datasets of urban acoustic scenes. To this end, we use a Gumbel-Softmax distribution as a soft approximation to the categorical distribution, allowing for training via backpropagation. In this settings, the softmax temperature serves as the main mechanism to tune clustering performance. The results show that the proposed model can obtain impressive clustering performance for all considered datasets, even when data points strongly overlap in time and frequency.

eess.AS↗

arXiv2026

Clustering of Acoustic Environments with Variational Autoencoders for Hearing Devices

Traditional acoustic environment classification relies on: i) classical signal processing algorithms, which are unable to extract meaningful representations of high-dimensional data; or on ii) supervised learning, limited by the availability of labels. Knowing that human-imposed labels do not always reflect the true structure of acoustic scenes, we explore the potential of (unsupervised) clustering of acoustic environments using variational autoencoders (VAEs). We employ a VAE model for categorical latent clustering with a Gumbel-Softmax reparameterization which can operate with a time-context windowing scheme for lower memory requirements, tailored for real-world hearing device scenarios. Additionally, general adaptations on VAE architectures for audio clustering are also proposed. The approaches are validated through the clustering of spoken digits, a simpler task where labels are meaningful, and urban soundscapes, where the recordings present strong overlap in time and frequency. While all variational methods succeeded when clustering spoken digits, only the proposed model achieved effective clustering performance on urban acoustic scenes, given its categorical nature.

eess.AS↗

arXiv2025

Hybrid Real- and Complex-valued Neural Network Architecture

We propose a \emph{hybrid} real- and complex-valued \emph{neural network} (HNN) architecture, designed to combine the computational efficiency of real-valued processing with the ability to effectively handle complex-valued data. We illustrate the limitations of using real-valued neural networks (RVNNs) for inherently complex-valued problems by showing how it learnt to perform complex-valued convolution, but with notable inefficiencies stemming from its real-valued constraints. To create the HNN, we propose to use building blocks containing both real- and complex-valued paths, where information between domains is exchanged through domain conversion functions. We also introduce novel complex-valued activation functions, with higher generalisation and parameterisation efficiency. HNN-specific architecture search techniques are described to navigate the larger solution space. Experiments with the AudioMNIST dataset demonstrate that the HNN reduces cross-entropy loss and consumes less parameters compared to an RVNN for all considered cases. Such results highlight the potential for the use of partially complex-valued processing in neural networks and applications for HNNs in many signal processing domains.

cs.LG↗

arXiv2025

Target Speaker Selection for Neural Network Beamforming in Multi-Speaker Scenarios

We propose a speaker selection mechanism (SSM) for the training of an end-to-end beamforming neural network, based on recent findings that a listener usually looks to the target speaker with a certain undershot angle. The mechanism allows the neural network model to learn toward which speaker to focus, during training, in a multi-speaker scenario, based on the position of listener and speakers. However, only audio information is necessary during inference. We perform acoustic simulations demonstrating the feasibility and performance when the SSM is employed in training. The results show significant increase in speech intelligibility, quality, and distortion metrics when compared to the minimum variance distortionless filter and the same neural network model trained without SSM. The success of the proposed method is a significant step forward toward the solution of the cocktail party problem.

eess.AS↗

arXiv2024

Spectral Masking with Explicit Time-Context Windowing for Neural Network-Based Monaural Speech Enhancement

We propose and analyze the use of an explicit time-context window for neural network-based spectral masking speech enhancement to leverage signal context dependencies between neighboring frames. In particular, we concentrate on soft masking and loss computed on the time-frequency representation of the reconstructed speech. We show that the application of a time-context windowing function at both input and output of the neural network model improves the soft mask estimation process by combining multiple estimates taken from different contexts. The proposed approach is only applied as post-optimization in inference mode, not requiring additional layers or special training for the neural network model. Our results show that the method consistently increases both intelligibility and signal quality of the denoised speech, as demonstrated for two classes of convolutional-based speech enhancement models. Importantly, the proposed method requires only a negligible ($\leq1\%$) increase in the number of model parameters, making it suitable for hardware-constrained applications.

eess.AS↗

arXiv2022

Systolic blood pressure estimation using ECG and PPG in patients undergoing surgery

Background and Objectives: In a significant portion of surgeries, blood pressure (BP) is often measured non-invasively in an intermittent manner. This practice has a risk of missing clinically relevant BP changes between two adjacent intermittent BP measurements. This study proposes a method to non-invasively estimate systolic blood pressure (SBP) with high accuracy in patients undergoing surgery. Methods: Continuous arterial BP, electrocardiography (ECG), and photoplethysmography (PPG) signals were acquired from 29 patients undergoing surgery. After extracting 9 features from the PPG and ECG signals, we dynamically selected features upon each intermittent measurement (every 10 min) of SBP based on feature robustness and the principle of correlation-based feature selection. Finally, multiple linear regression models were built to combine these features to estimate SBP every 30 s. Results: Compared to the reference SBP, the proposed method achieved a mean of difference at 0.08 mmHg, a standard deviation of difference at 7.97 mmHg, and a correlation coefficient at 0.89 (p < 0.001). Conclusions: This study demonstrates the feasibility of non-invasively estimating SBP every 30 s with high accuracy during surgery by using ECG, PPG, and intermittent SBP measurements every 10 min, which meets the standard of the Association for the Advancement of Medical Instrumentation. The proposed method has the potential to enhance BP monitoring in the operating room, improving patient outcomes and experiences.

physics.med-ph↗

arXiv2018

LSTM knowledge transfer for HRV-based sleep staging

Automated sleep stage classification using heart-rate variability is an active field of research. In this work limitations of the current state-of-the-art are addressed through the use of deep learning techniques and their efficacy is demonstrated. First, a temporal model is proposed for the inference of sleep stages from electrocardiography using a deep long- and short-term (LSTM) classifier and it is shown that this model outperforms previous approaches which were often limited to non-temporal or Markovian classifiers on a comprehensive benchmark data set (292 participants, 541214 samples) comprising a wide range of ages and pathological profiles, achieving a Cohen's $κ$ of $0.61\pm0.16$ and accuracy of $76.30\pm10.17$ annotated according to the Rechtschaffen & Kales annotation standard. Subsequently, it is demonstrated how knowledge learned on this large benchmark data set can be re-used through transfer learning for the classification of photoplethysmography (PPG) data. This is done using a smaller data set (60 participants, 91479 samples) that is annotated with the more recent American Association of Sleep Medicine annotation standard, achieving a Cohen's $κ$ of $0.63\pm0.13$ and accuracy of $74.65\pm8.63$ for wrist-mounted PPG-based sleep stage classification, higher than any previously reported performance using this sensor modality. This demonstrates the feasibility of knowledge transfer in sleep staging to adapt models for new sensor modalities as well as different annotation strategies.

q-bio.NC↗