DCASE 2022 Task 5: Few-shot Bioacoustic Event Detection Development Set

General Description:

The development set for task 5 of DCASE 2022 "Few-shot Bioacoustic Event Detection" consists of 209 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/

        |___JD/

            |____*.wav

            |____*.csv

        |___HT/

            |____*.wav

            |____*.csv

        |___BV/

            |____*.wav

            |____*.csv

        |___MT/

            |____*.wav

            |____*.csv

        |___WMW/

            |____*.wav

            |____*.csv

 

    |__Validation_Set/

        |___HB/

            |____*.wav

            |____*.csv

        |___PB/

            |____*.wav

            |____*.csv

        |___ME/

            |____*.wav

            |____*.csv

        |___ML/

            |____*.wav

            |____*.csv

 

Development_Set_AnnotationsOnly.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        |    174
Total duration                    |    21 hours
Total classes                    |    47
Total events                    |    14229
-----------------------------------------------------
TRAINING SET/BV
-----------------------------------------------------
Number of audio recordings        |    5
Total duration                    |    10 hours
Total classes                     |    11
Total events                     |    9026
Ratio event/duration            |    0.04
Sampling rate                    |    24000 Hz
-----------------------------------------------------
TRAINING SET/HT
-----------------------------------------------------
Number of audio recordings        |    5
Total duration                    |    5 hours
Total classes                     |    5
Total events                     |    611
Ratio event/duration            |    0.05
Sampling rate                    |    6000 Hz
-----------------------------------------------------
TRAINING SET/JD
-----------------------------------------------------
Number of audio recordings        |    1
Total duration                    |    10 mins
Total classes                    |    1
Total events                    |    357
Ratio event/duration            |    0.06
Sampling rate                    |    22050 Hz
-----------------------------------------------------
TRAINING SET/MT
-----------------------------------------------------
Number of audio recordings        |    2
Total duration                    |    1 hour and 10 mins
Total classes                    |    4
Total events                    |    1294
Ratio event/duration            |    0.04
Sampling rate                    |    8000 Hz
-----------------------------------------------------
TRAINING SET/WMW
-----------------------------------------------------
Number of audio recordings        |    161
Total duration                    |    4 hours and 40 mins
Total classes                    |    26
Total events                    |    2941
Ratio event/duration            |    0.24
Sampling rate                    |    various sampling rates
-----------------------------------------------------

-----------------------------------------------------
VALIDATION SET
-----------------------------------------------------
Number of audio recordings        |    35
Total duration                    |    6 hours and 17 minutes
Total classes                    |    22
Total events                     |    2007
-----------------------------------------------------
VALIDATION SET/HB
-----------------------------------------------------
Number of audio recordings        |    10
Total duration                    |    2 hours and 38 minutes
Total classes                     |    1
Total events                     |    607
Ratio event/duration            |    0.7
Sampling rate                    |    44100 Hz
-----------------------------------------------------
VALIDATION SET/PB
-----------------------------------------------------
Number of audio recordings        |    6
Total duration                    |    3 hours
Total classes                    |    2
Total events                     |    292
Ratio event/duration            |    0.003
Sampling rate                    |    44100 Hz
-----------------------------------------------------
VALIDATION SET/ME
-----------------------------------------------------
Number of audio recordings        |    2
Total duration                    |    20 minutes
Total classes                    |    2
Total events                     |    73
Ratio event/duration            |    0.01
Sampling rate                    |    48000 Hz
-----------------------------------------------------
VALIDATION SET/ML
-----------------------------------------------------
Number of audio recordings        |    17
Total duration                    |    20 minutes
Total classes                    |    17
Total events                     |    1035
Ratio event/duration            |    0.18
Sampling rate                    |    44100 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

DCASE2022_task5_training_set_classes.csv and DCASE2022_task5_validation_set_classes.csv provide a table with class code correspondence to class name for all classes in the Development set.

DCASE2022_task5_training_set_classes.csv
---------------------
dataset, class_code, class_name

DCASE2022_task5_validation_set_classes.csv
---------------------
dataset, recording, class_code, class_name

 

Evaluation Set

The Evaluation set for this task will be released on the 1st of June 2022

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:

Ines Nolasco -  i.dealmeidanolasco@qmul.ac.uk