Comparison of Frequency Cepstral Coefficients in Impulse Acoustic Events Detection

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F25%3A00369248" target="_blank" >RIV/68407700:21230/25:00369248 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1007/978-3-031-71397-2_1" target="_blank" >https://doi.org/10.1007/978-3-031-71397-2_1</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-031-71397-2_1" target="_blank" >10.1007/978-3-031-71397-2_1</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Comparison of Frequency Cepstral Coefficients in Impulse Acoustic Events Detection

Popis výsledku v původním jazyce

Various methods for acoustic impulse event detection and identification are available. They are usually based on time or frequency domain algorithms. Both these domains have their limitations and disadvantages. This article presents acoustic impulse events (such as gunshots) identification based on the Cepstral domain, combining the advantages of both frequency and time domains. It compares the efficiency of classification based on four different frequency Cepstral coefficients, namely Mel-frequency Cepstral Coefficients (MFCC), Inverse Mel-frequency Cepstral Coefficients (IMFCC), Linear-frequency Cepstral Coefficients (LFCC) and Gammatone-frequency Cepstral Coefficients (GFCC). These, originally speech features, showed to be promising in the other applications with good results. This work compares the classification accuracy of gunshots from several short and rifle guns and multiple impulse acoustic events (various types of slams, slaps, etc.) to represent false alarms. In total, more than four hundred acoustic event records have been acquired, where approx. 70% has been used for training, and the rest for validation. For a classification, a Support Vector Machine (SVM) classifier with 26 frequency Cepstral Coefficients from each MFCC, IMFCC, LFCC, and GFCC served as features are used. Accuracy and Matthew's correlation coefficient measure the classification success rate. The results confirm the superiority of GFCC to other analyzed methods.

Název v anglickém jazyce

Comparison of Frequency Cepstral Coefficients in Impulse Acoustic Events Detection

Popis výsledku anglicky

Various methods for acoustic impulse event detection and identification are available. They are usually based on time or frequency domain algorithms. Both these domains have their limitations and disadvantages. This article presents acoustic impulse events (such as gunshots) identification based on the Cepstral domain, combining the advantages of both frequency and time domains. It compares the efficiency of classification based on four different frequency Cepstral coefficients, namely Mel-frequency Cepstral Coefficients (MFCC), Inverse Mel-frequency Cepstral Coefficients (IMFCC), Linear-frequency Cepstral Coefficients (LFCC) and Gammatone-frequency Cepstral Coefficients (GFCC). These, originally speech features, showed to be promising in the other applications with good results. This work compares the classification accuracy of gunshots from several short and rifle guns and multiple impulse acoustic events (various types of slams, slaps, etc.) to represent false alarms. In total, more than four hundred acoustic event records have been acquired, where approx. 70% has been used for training, and the rest for validation. For a classification, a Support Vector Machine (SVM) classifier with 26 frequency Cepstral Coefficients from each MFCC, IMFCC, LFCC, and GFCC served as features are used. Accuracy and Matthew's correlation coefficient measure the classification success rate. The results confirm the superiority of GFCC to other analyzed methods.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20201 - Electrical and electronic engineering

Projekt

—

Návaznosti

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Modelling and Simulation for Autonomous Systems, 10th International Conference, MESAS 2023

ISBN

978-3-031-71396-5

ISSN

0302-9743

e-ISSN

1611-3349

Počet stran výsledku

8

Strana od-do

3-10

Název nakladatele

Springer, Cham

Místo vydání

—

Místo konání akce

Palermo

Datum konání akce

17. 10. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001419689800001