Sound Event Detection

Sound Event Detection (SED) is the task of recognizing the sound events and their respective temporal start and end time in a recording. Sound events in real life do not always occur in isolation, but tend to considerably overlap with each other. Recognizing such overlapping sound events is referred as polyphonic SED.

Source: A report on sound event detection with different binaural features

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 21–30 of 194 papers

Title	Date	Tasks	Status	Hype
MTDA-HSED: Mutual-Assistance Tuning and Dual-Branch Aggregating for Heterogeneous Sound Event Detection	Sep 10, 2024	Event DetectionSound Event Detection	CodeCode Available	0
From Computation to Consumption: Exploring the Compute-Energy Link for Training and Testing Neural Networks for SED Systems	Sep 8, 2024	Audio TaggingEvent Detection	—Unverified	0
Impact of Noisy Labels on Sound Event Detection: Deletion Errors Are More Detrimental Than Insertion Errors	Aug 27, 2024	Event DetectionSound Event Detection	—Unverified	0
MAT-SED: A Masked Audio Transformer with Masked-Reconstruction Based Pre-training for Sound Event Detection	Aug 16, 2024	Event DetectionSound Event Detection	CodeCode Available	2
SELD-Mamba: Selective State-Space Model for Sound Event Localization and Detection with Source Distance Estimation	Aug 9, 2024	Computational EfficiencyEvent Detection	—Unverified	0
Improving Audio Spectrogram Transformers for Sound Event Detection Through Multi-Stage Training	Jul 17, 2024	Event DetectionMissing Labels	CodeCode Available	1
Multi-Iteration Multi-Stage Fine-Tuning of Transformers for Sound Event Detection with Heterogeneous Datasets	Jul 17, 2024	Event DetectionSound Event Detection	CodeCode Available	1
UCIL: An Unsupervised Class Incremental Learning Approach for Sound Event Detection	Jul 4, 2024	class-incremental learningClass Incremental Learning	—Unverified	0
Mixstyle based Domain Generalization for Sound Event Detection with Heterogeneous Training Data	Jul 4, 2024	Domain GeneralizationEvent Detection	—Unverified	0
WildDESED: An LLM-Powered Dataset for Wild Domestic Environment Sound Event Detection System	Jul 4, 2024	Event DetectionLanguage Modeling	CodeCode Available	1

Show:10 25 50

← PrevPage 3 of 20Next →

All datasets DESED L3DAS21 WildDESED Mivia Audio Events Mivia Road Events

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	ATST-SED	event-based F1 score	63.4	—	Unverified
2	SE-CRNN-16 with DualKD	event-based F1 score	55.6	—	Unverified
3	FDY-CRNN	event-based F1 score	54	—	Unverified
4	HTS-AT	event-based F1 score	50.7	—	Unverified
5	RCT	event-based F1 score	49.62	—	Unverified
6	FiltAug SED	event-based F1 score	49.6	—	Unverified
7	SED-SSep baseline dcase task 4 2020 v2	event-based F1 score	40.7	—	Unverified
8	Baseline dcase task 4 2020 v2	event-based F1 score	39	—	Unverified
9	Baseline	event-based F1 score	25.8	—	Unverified
10	MAT-SED	PSDS1	0.59	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	PHC SEDnet n=8	Error Rate	0.56	—	Unverified
2	Quaternion SEDnet	Error Rate	0.52	—	Unverified
3	PHC SEDnet n=16	Error Rate	0.51	—	Unverified
4	PHC SEDnet n=4	Error Rate	0.45	—	Unverified
5	PHC SEDnet n=2	Error Rate	0.39	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	CRNN (with BEATs + Separation)	PSDS1 (-5dB)	0.13	—	Unverified
2	CRNN (with BEATs)	PSDS1 (-5dB)	0.07	—	Unverified
3	CRNN (WildDESED + Curriculrm learning)	PSDS1 (-5dB)	0.05	—	Unverified
4	CRNN (WildDESED)	PSDS1 (-5dB)	0.05	—	Unverified
5	CRNN	PSDS1 (-5dB)	0.02	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	DENet	Rank-1 Recognition Rate	0.98	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	DENet	Rank-1 Recognition Rate	1	—	Unverified