Domestic environment sound event detection
The DESED dataset is a dataset designed to recognize sound event classes in domestic environments. The dataset is designed to be used for sound event detection (SED, recognize events with their time boundaries) but it can also be used for sound event tagging (SET, indicate presence of an event in an audio file).
The dataset is composed of 10 event classes to recognize in 10 second audio files. The classes are: Alarm/bell/ringing, Blender, Cat, Dog, Dishes,
Electric shaver/toothbrush, Frying, Running water, Speech, Vacuum cleaner.
Source: https://project.inria.fr/desed/
Image Source: https://project.inria.fr/desed/
Variants: DESED
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Sound Event Detection | JiTTER | JiTTER: Jigsaw Temporal Transformer for … | 2025-02-28 |
| Sound Event Detection | MAT-SED | MAT-SED: A Masked Audio Transformer … | 2024-08-16 |
| Sound Event Detection | ABC + MDFD-CRNN | Pushing the Limit of Sound … | 2024-06-19 |
| Sound Event Detection | MDFD-CRNN | Pushing the Limit of Sound … | 2024-06-19 |
| Sound Event Detection | SE-CRNN-16 with DualKD | Dual Knowledge Distillation for Efficient … | 2024-02-05 |
| Sound Event Detection | ATST-SED | Fine-tune the pretrained ATST model … | 2023-09-15 |
| Sound Event Detection | FDY-CRNN | Frequency Dynamic Convolution: Frequency-Adaptive Pattern … | 2022-03-29 |
| Sound Event Detection | HTS-AT | HTS-AT: A Hierarchical Token-Semantic Audio … | 2022-02-02 |
| Sound Event Detection | RCT | RCT: Random Consistency Training for … | 2021-10-21 |
| Sound Event Detection | FiltAug SED | Heavily Augmented Sound Event Detection … | 2021-07-08 |
Recent papers with results on this dataset: