# DCASE2019 Challenge - Task1B Leaderboard — Data
来源：https://www.kaggle.com/competitions/dcase2019-task1b-leaderboard/data
采集时间：2026-10-20T13:07:54Z
The data for the competition consists of **training dataset**, 10-second audio segments released together with the scene class labels, and **test dataset** which is released without the scene labels. Your task is to predict the scene labels for the audio segments in the test dataset.
# Datasets
## Training dataset
The development dataset provided with DCASE2019 Challenge task 1B is used as training dataset. There are three methods for obtaining the training dataset.
1) Dataset can be downloaded directly
***
### [TAU Urban Acoustic Scenes 2019 Mobile, Development dataset (18.7 GB) ](https://zenodo.org/record/2589332)
[](https://doi.org/10.5281/zenodo.2589332)
***
2) Dataset can be downloaded using the provided download script:
`python task1b_dataset_downloader.py -d train`
3) If you are using official [DCASE2019 Challenge Task 1 baseline system](https://github.com/toni-heittola/dcase2019_task1_baseline), you can use it to download the dataset too:
`python task1b.py -m dev --download_dataset datasets/`
### Organization of the dataset
The dataset package contains readme file, which explains how the dataset is organized. Reference data for the provided audio segments can be found in CSV-formated file `meta.csv`:
filename scene_label identifier source_label
audio/airport-barcelona-0-0-a.wav airport barcelona-0 a
audio/airport-barcelona-0-1-a.wav airport barcelona-0 a
....
audio/bus-barcelona-17-706-a.wav bus barcelona-17 a
audio/bus-helsinki-18-707-a.wav bus helsinki-18 a
audio/bus-helsinki-18-708-a.wav bus helsinki-18 a
...
audio/metro-london-46-1402-b.wav metro london-46 b
audio/metro-stockholm-57-1699-b.wav metro stockholm-57 b
...
audio/street_traffic-vienna-178-5490-c.wav street_traffic vienna-178 c
audio/tram-helsinki-182-5639-c.wav tram helsinki-182 c
...
audio/tram-vienna-285-8637-a.wav tram vienna-285 a
audio/tram-vienna-285-8638-a.wav tram vienna-285 a
audio/tram-vienna-285-8639-a.wav tram vienna-285 a
Naming convention for audio segments is file naming convention: [scene label]-[city]-[location id]-[segment id]-[device id].wav
`source_label` field in the meta data or `[device id]` field in the audio segment filename is used to denote the used recording device. In this competition, all material is captured with the same device (`a`).
The dataset is provided with single fold cross-validation set (training / testing) to ease the usage of the development dataset. Participants can do their own cross-validation setup too. File lists assigning segments to the subsets are provided under directory `evaluation_setup`.
## Testing dataset
Similarly to the development dataset, there are three methods for obtaining the dataset.
1) Downloading it directly:
***
### [TAU Urban Acoustic Scenes 2019 Mobile, Leaderboard dataset (1.3 GB) ](https://zenodo.org/record/2673004)
[](https://doi.org/10.5281/zenodo.2673004)
***
2) Dataset can be downloaded using the provided download script:
`python task1b_dataset_downloader.py -d test`
3) If you are using official [DCASE2019 Challenge Task 1 baseline system](https://github.com/toni-heittola/dcase2019_task1_baseline) you can use it to download the dataset too:
`python task1b.py -m leaderboard --download_dataset datasets/`
## Dataset license
License permits free academic usage. Any commercial use is strictly prohibited. For commercial use, contact dataset authors.
Copyright (c) 2019 Tampere University and its licensors
All rights reserved.
Permission is hereby granted, without written agreement and without license or royalty
fees, to use and copy the TAU Urban Acoustic Scenes 2019 (“Work”) described in this document
and composed of audio and metadata. This grant is only for experimental and non-commercial
purposes, provided that the copyright notice in its entirety appear in all copies of this Work,
and the original source of this Work, (Audio Research Group at Tampere University of Technology),
is acknowledged in any publication that reports research using this Work.
Any commercial use of the Work or any part thereof is strictly prohibited.
Commercial use include, but is not limited to:
- selling or reproducing the Work
- selling or distributing the results or content achieved by use of the Work
- providing services by using the Work.
IN NO EVENT SHALL TAMPERE UNIVERSITY OR ITS LICENSORS BE LIABLE TO ANY PARTY
FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE
OF THIS WORK AND ITS DOCUMENTATION, EVEN IF TAMPERE UNIVERSITY OR ITS
LICENSORS HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
TAMPERE UNIVERSITY AND ALL ITS LICENSORS SPECIFICALLY DISCLAIMS ANY
WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE. THE WORK PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND
THE TAMPERE UNIVERSITY HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT,
UPDATES, ENHANCEMENTS, OR MODIFICATIONS.