Sound Event Detection in Domestic Environments with Weakly Labeled Data and Soundscape Synthesis

By Nicolas Turpault, Romain Serizel, Ankit Shah, Justin Salamon. 2019

340 citations recorded by the source corpus.

This paper presents Task 4 of the Detection and Classification of Acoustic Scenes and Events (DCASE) 2019 challenge and provides a first analysis of the challenge results. The task is a followup to Task 4 of DCASE 2018, and involves training systems for large-scale detection of sound events using a combination of weakly labeled data, i.e. training labels without time boundaries, and strongly-labeled synthesized data. The paper introduces Domestic Environment Sound Event Detection (DESED) dataset mixing a part of last year dataset and an additional synthetic, strongly labeled, dataset provided this year that we’ll describe more in detail. We also report the performance of the submitted systems on the official evaluation (test) and development sets as well as several additional datasets. The best systems from this year outperform last year’s winning system by about 10% points in terms of F-measure.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

FSD50K: An Open Dataset of Human-Labeled Sound Events (2020, background): “An increasingly popular dataset is DESED [28], covering 10 classes of domestic sounds.”
Limitations of Weak Labels for Embedding and Tagging (2020, positive): “This approach is directly inspired by the DESED synthetic dataset [23].”
The Benefit of Temporally-Strong Labels in Audio Event Classification (2021, methodology): “DESED [7] has been used for DCASE Task 4 in recent years, covering 10 classes of domestic sounds.”
Convolutional Recurrent Neural Network and Data Augmentation for Audio Tagging with Noisy Labels and Minimal Supervision (2019, background): “HA 3455/15-1 within the Research Unit FOR 2457 (acoustic sensor networks). semi-supervised approaches which directly beneﬁt from unlabeled data [3, 4], which is abundant.”
Sound Event Detection in Synthetic Domestic Environments (2020, neutral): “The DESED 1 dataset is composed of 10-sec audio clips recorded in a domestic environment [7, 10] or synthesized to simulate a domestic environment.”
Improving Sound Event Detection in Domestic Environments using Sound Separation (2020, methodology): “The dataset used for the SED experiments is DESED(1), a flexible dataset for SED in domestic environments composed of 10-sec audio clips that are recorded or synthesized [13, 4].”