Publications associées à Blind Audiovisual Separation based on Redundant Representations

Learning Sparse Graphons And The Generalized Kesten-Stigum Threshold

The problem of learning graphons has attracted considerable attention across several scientific communities, with significant progress over the re-cent years in sparser regimes. Yet, the current techniques still require diverg-ing degrees in order to succe ...

INST MATHEMATICAL STATISTICS-IMS2023

Learnable filter-banks for CNN-based audio applications

Nicolas Aspert, Benjamin Ricaud, Laurent Colbois, Helena Peic Tukuljac

We investigate the design of a convolutional layer where kernels are parameterized functions. This layer aims at being the input layer of convolutional neural networks for audio applications or applications involving time-series. The kernels are defined as ...

2022

Unsupervised Voice Activity Detection by Modeling Source and System Information using Zero Frequency Filtering

Mathew Magimai Doss, Eklavya Sarkar

Voice activity detection (VAD) is an important pre-processing step for speech technology applications. The task consists of deriving segment boundaries of audio signals which contain voicing information. In recent years, it has been shown that voice source ...

ISCA2022

Concurrent Evolution of Biomechanical and Physiological Parameters With Running-Induced Acute Fatigue

Kamiar Aminian, Anisoara Ionescu, Gaëlle Prigent, Salil Apte, Vincent Gremeaux

Understanding the influence of running-induced acute fatigue on the homeostasis of the body is essential to mitigate the adverse effects and optimize positive adaptations to training. Fatigue is a multifactorial phenomenon, which influences biomechanical, ...

FRONTIERS MEDIA SA2022

Practical Compilation of Quantum Programs

Bruno Schmitt Antunes

It's been a little more than 40 years since researchers first suggested exploiting quantum physics to build more powerful computers. During this time, we have seen the development of many quantum algorithms and significant technological advances to make th ...

EPFL2022

Radical Intangibles: Materializing the Ephemeral

Sarah Irene Brutton Kenderdine, Lillian Hibberd, Jeffrey Shaw

New materialism considers that the world and its histories are produced by a range of material forces that extend from the physical and the biological to the psychological, social and cultural. In recognizing that heritage is not held in objects alone, new ...

2021

Multi-task Neural Network for Robust Multiple Speaker Embedding Extraction

Jean-Marc Odobez, Petr Motlicek, Weipeng He

This paper introduces a novel approach for extracting speaker embeddings from audio mixtures of multiple overlapping voices. This approach is based on a multi-task neural network. The network first extracts a latent feature for each direction. This feature ...

ISCA-INT SPEECH COMMUNICATION ASSOC2021

ASAP: a Dataset of Aligned Scores and Performances for Piano Transcription

Andrew Philip McLeod

In this paper we present Aligned Scores and Performances (ASAP): a new dataset of 222 digital musical scores aligned with 1068 performances (more than 92 hours) of Western classical piano music.The scores are provided as paired MusicXML files and quantized ...

2020

Sparse and Parametric Modeling with Applications to Acoustics and Audio

Helena Peic Tukuljac

Recent advances in signal processing, machine learning and deep learning with sparse intrinsic structure of data have paved the path for solving inverse problems in acoustics and audio. The main task of this thesis was to bridge the gap between the powerfu ...

EPFL2020

Multisensory haptic system and method

Olaf Blanke, Simon Gallo, Giulio Rognini

A computer-implemented method for operating a haptic device, the haptic device comprising a plurality of tactile displays configured to provide haptic stimuli to a user, the method including the steps of (a) processing an audio signal derived from an audio ...

2020

Inpainting of Long Audio Segments With Similarity Graphs

Nathanaël Perraudin

We present a novel method for the compensation of long duration data loss in audio signals, in particular music. The concealment of such signal defects is based on a graph that encodes signal structure in terms of time-persistent spectral similarity. A sui ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2018

Visual speech recognition : from traditional to deep learning frameworks

Marina Zimmermann

Speech is the most natural means of communication for humans. Therefore, since the beginning of computers it has been a goal to interact with machines via speech. While there have been gradual improvements in this field over the decades, and with recent dr ...

EPFL2018

Deep Learning of Human Perception in Audio Event Classification

Samuel Denys Beuret, Yi Yu

In this paper, we introduce our recent studies on human perception in audio event classification. In particular, the pre-trained model VGGish is used as feature extractor to process audio data, and DenseNet is trained by and used as feature extractor for o ...

IEEE2018

Experimental evaluation of speech enhancement methods in remote microphone systems for hearing aids

Hervé Lissek, Gilles André Courtois, Vincent Pierre Olivier Grimaldi, Ina Kodrasi

Enhancing speech intelligibility for hearing-impaired subjects in complex acoustic conditions is still a challenging topic of research. To mitigate the detrimental effects of background noise and reverberation, current hearing instruments incorporate vario ...

European Acoustics Association2018

Near-Optimal Noisy Group Testing via Separate Decoding of Items

Volkan Cevher, Jonathan Mark Scarlett

The group testing problem consists of determining a small set of defective items from a larger set of items based on a number of tests, and is relevant in applications such as medical testing, communication protocols, pattern matching, and more. In this pa ...

2018

Rake, Peel, Sketch : The Signal Processing Pipeline Revisited

Robin Scheibler

The prototypical signal processing pipeline can be divided into four blocks. Representation of the signal in a basis suitable for processing. Enhancement of the meaningful part of the signal and noise reduction. Estimation of important statistical properti ...

EPFL2017

Method and apparatus for presenting to a user of a wearable apparatus additional information related to an audio scene

Martin Vetterli, Juri Ranieri, Ivan Dokmanic

A method for presenting to a user of a wearable audio device a modified audio scene together with additional information related to the audio scene, comprising: capturing audio signals with a plurality of microphones; outputting an audio signal with a plur ...

2017

Jump-Penalized Least Absolute Values Estimation of Scalar or Circle-Valued Signals

Michaël Unser, Martin Kurt Storath

We study jump-penalized estimators based on least absolute deviations which are often referred to as Potts estimators. They are estimators for a parsimonious piecewise constant representation of noisy data having a noise distribution which has heavier tail ...

Oxford University Press2017

System fusion and speaker linking for longitudinal diarization of TV shows

Hervé Bourlard, Petr Motlicek

Performing speaker diarization while uniquely identifying the speakers in a collection of audio recordings is a challenging task. Based on our previous work on speaker diarization and linking, we developed a system for diarizing longitudinal TV show data s ...

IEEE2016

Efficient Algorithms for the Data Exchange Problem

Michael Christoph Gastpar, Sameer Pawar

In this paper, we study the data exchange problem, where a set of users is interested in gaining access to a common file, but where each has only partial knowledge about it as side-information. Assuming that the file is broken into packets, the side-inform ...

Institute of Electrical and Electronics Engineers2016