Tampere University of Technology

TUTCRIS Research Portal

Metrics for polyphonic sound event detection

Research output: Contribution to journalArticleScientificpeer-review

Standard

Metrics for polyphonic sound event detection. / Mesaros, Annamaria; Heittola, Toni; Virtanen, Tuomas.

In: Applied Sciences, Vol. 6, No. 6, 162, 2016.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

APA

Vancouver

Author

Bibtex - Download

@article{6a13d4fa99f54f3fb088908ea1af9ef8,
title = "Metrics for polyphonic sound event detection",
abstract = "This paper presents and discusses various metrics proposed for evaluation of polyphonic sound event detection systems used in realistic situations where there are typically multiple sound sources active simultaneously. The system output in this case contains overlapping events, marked as multiple sounds detected as being active at the same time. The polyphonic system output requires a suitable procedure for evaluation against a reference. Metrics from neighboring fields such as speech recognition and speaker diarization can be used, but they need to be partially redefined to deal with the overlapping events. We present a review of the most common metrics in the field and the way they are adapted and interpreted in the polyphonic case. We discuss segment-based and event-based definitions of each metric and explain the consequences of instance-based and class-based averaging using a case study. In parallel, we provide a toolbox containing implementations of presented metrics.",
keywords = "Audio content analysis, Audio signal processing, Computational auditory scene analysis, Evaluation of sound event detection, Everyday sounds, Pattern recognition, Polyphonic sound event detection, Sound events",
author = "Annamaria Mesaros and Toni Heittola and Tuomas Virtanen",
year = "2016",
doi = "10.3390/app6060162",
language = "English",
volume = "6",
journal = "Applied Sciences",
issn = "2076-3417",
publisher = "MDPI",
number = "6",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Metrics for polyphonic sound event detection

AU - Mesaros, Annamaria

AU - Heittola, Toni

AU - Virtanen, Tuomas

PY - 2016

Y1 - 2016

N2 - This paper presents and discusses various metrics proposed for evaluation of polyphonic sound event detection systems used in realistic situations where there are typically multiple sound sources active simultaneously. The system output in this case contains overlapping events, marked as multiple sounds detected as being active at the same time. The polyphonic system output requires a suitable procedure for evaluation against a reference. Metrics from neighboring fields such as speech recognition and speaker diarization can be used, but they need to be partially redefined to deal with the overlapping events. We present a review of the most common metrics in the field and the way they are adapted and interpreted in the polyphonic case. We discuss segment-based and event-based definitions of each metric and explain the consequences of instance-based and class-based averaging using a case study. In parallel, we provide a toolbox containing implementations of presented metrics.

AB - This paper presents and discusses various metrics proposed for evaluation of polyphonic sound event detection systems used in realistic situations where there are typically multiple sound sources active simultaneously. The system output in this case contains overlapping events, marked as multiple sounds detected as being active at the same time. The polyphonic system output requires a suitable procedure for evaluation against a reference. Metrics from neighboring fields such as speech recognition and speaker diarization can be used, but they need to be partially redefined to deal with the overlapping events. We present a review of the most common metrics in the field and the way they are adapted and interpreted in the polyphonic case. We discuss segment-based and event-based definitions of each metric and explain the consequences of instance-based and class-based averaging using a case study. In parallel, we provide a toolbox containing implementations of presented metrics.

KW - Audio content analysis

KW - Audio signal processing

KW - Computational auditory scene analysis

KW - Evaluation of sound event detection

KW - Everyday sounds

KW - Pattern recognition

KW - Polyphonic sound event detection

KW - Sound events

U2 - 10.3390/app6060162

DO - 10.3390/app6060162

M3 - Article

VL - 6

JO - Applied Sciences

JF - Applied Sciences

SN - 2076-3417

IS - 6

M1 - 162

ER -