Voice health is traditionally assessed by methods that rely on the perception of a clinician, who integrates auditory and visual cues in order to reach a conclusion about the voice under evaluation. However, these tasks suffer from inter-professional variability due to its subjective nature, which is why more objective, computational-based methods are of interest. Two examples of such subjective tasks are the classification of voices in three types according to their periodicity, also termed voice typing, and the evaluation of six aspects of voice quality by means of the consensus auditory-perceptual evaluation of voice (CAPE-V) protocol. In this paper, two approaches to emulate each of those tasks are introduced, based on simple features extracted from the second-order scattering transform coefficients and support vector machines. Firstly, a system for automatic voice typing was trained and its classification performance was evaluated for intra and inter-dataset trials using two widely known corpora. Accuracies above 80%, comparable to the state-of-the-art, were found for all the experiments conducted. Secondly, a multidimensional, multioutput regression chain model was used to automatically grade the voice quality features of the CAPE-V protocol, obtaining errors and correlation coefficients that are comparable to those found for three human raters.

Voice jitter is defined as a random perturbation of the glottal cycle duration which can be useful for voice parametrization and that usually depends on finding fiducial points in this signal. In this paper, a novel application of the Fourier-based high-order synchrosqueezing (FSSTN) operators on voice signals is introduced for voice jitter estimation without period-segmentation. To this end, an innovative interpretation of the relative jitter formula in terms of the total variation of the sequence of periods is proposed. This allows us to derive an algorithm for jitter estimation that uses the (continuous) instantaneous fundamental frequency of the signal and its first derivative (chirp rate) which can be obtained, respectively, from the local complex frequency and the local complex modulation FSSTN operators. Numerical experiments using synthetic signals with known true jitter show that this novel approach yields similar results to other state-of-the-art method, PRAAT, for true jitter within the range [0.2%, 1.2%], and that it outperforms PRAAT for true jitter values in the range [1%, 15%]. The here proposed method seems a promising tool for voice jitter estimation and constitutes a novel application of the high-order synchrosqueezing operators for voice signals with potential impact on jitter modeling and on the clinical field.

This paper introduces a 2D extension of the empirical mode decomposition (EMD), through a novel approach based on unconstrained optimization. EMD is a fully data-driven method that locally separates, in a completely data-driven and unsupervised manner, signals into fast and slow oscillations. The present proposal implements the method in a very simple and fast way, and it is compared with the state-of-the-art methods evidencing the advantages of being computationally efficient, orientation-independent, and leads to better performances for the decomposition of amplitude modulated-frequency modulated (AM-FM) images. The resulting genuine 2D method is successfully tested on artificial AM-FM images and its capabilities are illustrated on a biomedical example. The proposed framework leaves room for an nD extension (n > 2 ).

In this paper an algorithm based on the ensemble empirical mode decomposition (EEMD) is presented. The key idea on the EEMD relies on averaging the modes obtained by EMD applied to several realizations of Gaussian white noise added to the original signal. The resulting decomposition solves the EMD mode mixing problem, however it introduces new ones. In the method here proposed, a particular noise is added at each stage of the decomposition and a unique residue is computed to obtain each mode. The resulting decomposition is complete, with a numerically negligible error. Two examples are presented: a discrete Dirac delta function and an electrocardiogram signal. The results show that, compared with EEMD, the new method here presented also provides a better spectral separation of the modes and a lesser number of sifting iterations is needed, reducing the computational cost.