Sounding Industry: Challenges and Datasets for Industrial Sound Analysis

Grollmisch, Sascha; Abeber, Jakob; Liebetrau, Judith; Lukashevich, Hanna

doi:10.23919/eusipco.2019.8902941

Cited by 36 publications

(17 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many challenges arise when ASC models are deployed in smart city [90,91] or industrial sound analysis [92] scenarios. The first challenge is the model complexity, which is limited if data privacy concerns require the classification to be performed directly on mobile sensor devices.…”

Section: Real-world Deploymentmentioning

confidence: 99%

A Review of Deep Learning Based Methods for Acoustic Scene Classification

Abeßer

2020

Applied Sciences

129

View full text Add to dashboard Cite

The number of publications on acoustic scene classification (ASC) in environmental audio recordings has constantly increased over the last few years. This was mainly stimulated by the annual Detection and Classification of Acoustic Scenes and Events (DCASE) competition with its first edition in 2013. All competitions so far involved one or multiple ASC tasks. With a focus on deep learning based ASC algorithms, this article summarizes and groups existing approaches for data preparation, i.e., feature representations, feature pre-processing, and data augmentation, and for data modeling, i.e., neural network architectures and learning paradigms. Finally, the paper discusses current algorithmic limitations and open challenges in order to preview possible future developments towards the real-life application of ASC systems.

show abstract

Section: Real-world Deploymentmentioning

confidence: 99%

A Review of Deep Learning Based Methods for Acoustic Scene Classification

Abeßer

2020

Applied Sciences

129

View full text Add to dashboard Cite

show abstract

“…The 10 best performing submissions in 2020 applied Deep Learning (DL), treating the developement dataset as training data. Some used different forms of data augmentation and additions from ex-ternal datasets such as AudioSet [6] and Fraunhofer's IDMT-ISA-EE dataset [2]. For inference, most submissions directly applied the trained DL models, often via multi-task ensembles.…”

Section: Uad In Dcasementioning

confidence: 99%

“…For our data sources, we merged the Development and Additional Training datasets [12] from DCASE 2021, task 2. To illustrate the role of external datasets, we also incorporated the 10-second cut variant of Fraunhofer's IDMT-ISA-EE dataset [2], and a custom subset of AudioSet consisting of 10-second segments from ∼40k unique videos. All audio files were converted to mono 16kHz, and (−1, 1) normalization was applied.…”

Section: Umap For Dcase 2021mentioning

confidence: 99%

“…This can result in false negatives, if the detector is not sensitive enough, or false positives, if the detector does not tolerate or adapt to domain shifts. In the audio domain, UAD has attracted attention as a promising Predictive Maintenance (PdM) [1] solution for Industrial Sound Analysis (ISA) [2]: Sound monitoring is non-invasive, is robust to occlusions, can be carried out during production, and anomalous sounds can signal issues long before critical faults occur. UAD-S is a natural extension to UAD, since even in controlled environments, 1 DCASE website: http://dcase.community 2 Online Resources: Code: https://github.com/andres-fr/dcase2021_umaps Webpage: https://ai4s.surrey.ac.uk/2021/dcase_uads new non-anomalous sources of sound can arise (e.g.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using Umap to Inspect Audio Data for Unsupervised Anomaly Detection Under Domain-Shift Conditions

Fernandez¹,

Plumbley²

2021

Preprint

View full text Add to dashboard Cite

The goal of \ac{UAD} is to detect anomalous signals under the condition that only non-anomalous ({\it normal}) data is available beforehand. In UAD under Domain-Shift Conditions (UAD-S), data is further exposed to contextual changes that are usually unknown beforehand. Motivated by the difficulties encountered in the \acs{UAD-S} task presented at the 2021 edition of the \ac{DCASE} challenge\footnote{DCASE website: \url{http://dcase.community}}, we visually inspect \acp{UMAP} for log-STFT, log-mel and pretrained \ac{L3} representations of the \ac{DCASE} \acs{UAD-S} dataset. In our exploratory investigation, we look for two qualities, {\it \ac{SEP}} and {\it \ac{DSUP}}, and formulate several hypotheses that could facilitate diagnosis and developement of further representation and detection approaches. Particularly, we hypothesize that input length and pretraining may regulate a relevant tradeoff between \ac{SEP} and \ac{DSUP}. Our code as well as the resulting \acp{UMAP} and plots are publicly available\footnote{Online Resources:\\ Code: \url{https://github.com/andres-fr/dcase2021_umaps}\\Webpage: \url{https://ai4s.surrey.ac.uk/2021/dcase_uads}}.

show abstract

“…Finally, sound requires lesser computational resources as compared to an image or video [3], [4]. Thus, a SED system has great potential in many other domains besides speech recognition [5], [6] such as medical telemonitoring [7], surveillance [8], [9], equipment monitoring [10] [11] and wildlife monitoring [12], [13].…”

Section: Introductionmentioning

confidence: 99%

A Comprehensive Review of Polyphonic Sound Event Detection

Chan

Chin

2020

IEEE Access

View full text Add to dashboard Cite

One of the most amazing functions of the human auditory system is the ability to detect all kinds of sound events in the environment. With the technologies and hardware advances, polyphonic Sound Event Detection (SED) can be developed to mimic the ability of the human auditory system. However, the development of a SED system is no trivial task, and several different factors often hinder accuracy. Although there are several overview papers available, most of them only provide a theoretical overview of algorithms used with little discussion. Thus, to the best of the authors' knowledge, there is no comprehensive review that covers this particular domain. Therefore, this paper aims to provide an in-depth discussion of different methodologies proposed by various authors that include the features used, detection algorithms, and their corresponding accuracy and limitations. Additional information on possible trends is also discussed that can be useful for future development works.

show abstract

Sounding Industry: Challenges and Datasets for Industrial Sound Analysis

Cited by 36 publications

References 11 publications

A Review of Deep Learning Based Methods for Acoustic Scene Classification

A Review of Deep Learning Based Methods for Acoustic Scene Classification

Using Umap to Inspect Audio Data for Unsupervised Anomaly Detection Under Domain-Shift Conditions

A Comprehensive Review of Polyphonic Sound Event Detection

Contact Info

Product

Resources

About