DCASE 2021 Task 5: Few-shot Bioacoustic Event Detection Development Set

General Description

The development set for task 5 of DCASE 2021 "Few-shot Bioacoustic Event Detection" consists of 19 audio files acquired from different bioacoustic sources. The dataset is split into training and validation Sets. 

Multi-class annotations are provided for the training set with positive (POS), negative (NEG) and unkwown (UNK) values for each class. UNK indicates uncertainty about a class. 

Single-class (class of interest) annotations are provided for the validation set, with events marked as positive (POS) or unkwown (UNK) provided for the class of interest. 

 

Folder Structure

Development_Set.zip

|_Development_Set/

    |__Training_Set/

        |___BV/

            |____*.wav

            |____*.csv

        |___HT/

            |____*.wav

            |____*.csv

        |___JD/

            |____*.wav

            |____*.csv

        |___MT/

            |____*.wav

            |____*.csv

    |__Validation_Set/

        |___HV/

            |____*.wav

            |____*.csv

        |___PB/

            |____*.wav

            |____*.csv

 

Development_Set_Audio.zip has the same structure but contains only the *.wav files.

Development_Set_Annotations.zip has the same structure but contains only the *.csv files

 

Dataset statistics

Some statistics on this dataset are as follows, split between training and validation set and their sub-folders:

-----------------------------------------------------
TRAINING SET
-----------------------------------------------------
Number of audio recordings        |    11
Total duration                    |    14 hours and 20 mins
Total classes (excl. UNK)        |    19
Total events (excl. UNK)        |    4,686
-----------------------------------------------------
TRAINING SET/BV
-----------------------------------------------------
Number of audio recordings        |    5
Total duration                    |    10 hours
Total classes (excl. UNK)        |    11
Total events (excl. UNK)        |    2,662
Sampling rate                    |    24,000 Hz
-----------------------------------------------------
TRAINING SET/HT
-----------------------------------------------------
Number of audio recordings        |    3
Total duration                    |    3 hours
Total classes (excl. UNK)        |    3
Total events (excl. UNK)        |    435
Sampling rate                    |    6,000 Hz
-----------------------------------------------------
TRAINING SET/JD
-----------------------------------------------------
Number of audio recordings        |    1
Total duration                    |    10 mins
Total classes (excl. UNK)        |    1
Total events (excl. UNK)        |    355
Sampling rate                    |    22,050 Hz
-----------------------------------------------------
TRAINING SET/MT
-----------------------------------------------------
Number of audio recordings        |    2
Total duration                    |    1 hour and 10 mins
Total classes (excl. UNK)        |    4
Total events (excl. UNK)        |    1,234
Sampling rate                    |    8,000 Hz
-----------------------------------------------------

-----------------------------------------------------
VALIDATION SET
-----------------------------------------------------
Number of audio recordings        |    8
Total duration                    |    5 hours
Total classes (excl. UNK)        |    4
Total events (excl. UNK)        |    310
-----------------------------------------------------
VALIDATION SET/HV
-----------------------------------------------------
Number of audio recordings        |    2
Total duration                    |    2 hours
Total classes (excl. UNK)        |    2
Total events (excl. UNK)        |    50
Sampling rate                    |    6,000 Hz
-----------------------------------------------------
VALIDATION SET/PB
-----------------------------------------------------
Number of audio recordings        |    6
Total duration                    |    3 hours
Total classes (excl. UNK)        |    2
Total events (excl. UNK)        |    260
Sampling rate                    |    44,100 Hz
-----------------------------------------------------

 

Annotation structure

Each line of the annotation csv represents an event in the audio file. The column descriptions are as follows:

TRAINING SET
---------------------
Audiofilename, Starttime, Endtime, CLASS_1, CLASS_2, ...CLASS_N

VALIDATION SET
---------------------
Audiofilename, Starttime, Endtime, Q

 

Classes

DCASE2021_task5_training_set_classes.csv and DCASE2021_task5_validation_set_classes.csv provide a table with class code correspondace to class name for all classes in the Development set.

DCASE2021_task5_training_set_classes.csv
---------------------
dataset, class_code, class_name

DCASE2021_task5_validation_set_classes.csv
---------------------
dataset, recording, class_code, class_name

 

Evaluation Set

The Evaluation set for the same task can be found at: https://doi.org/10.5281/zenodo.5413149

 

Open Access

This dataset is available under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.

Contact info

Please send any feedback or questions to:
Veronica Morfi: g.v.morfi@qmul.ac.uk