DCASE 2018 task3 Bird Audio Detection介绍

DCASE是Detection and Classification of Acoustic Scenes and Events的缩写,翻译过来就是声学场景与事件的检测与分类[1]。DCASE含有竞赛(challenge)和研讨会(workshop)两部分。该竞赛的目的在于提供声学场景和事件的公开数据集,鼓励大家使用和比较不同的方法的优缺点,促进声学场景和事件领域的向前发展。分别在2013、2016、2017年举办过DCASE Challenge比赛,今年是DCASE Challenge 2018,包含五个任务,本文主要讲解的是任务3(task3),该任务是Bird Aduio Detection(鸟叫声检测,以下简称BAD)[2]。

1. 描述

给定三个开发集,三个评估集。三个开发集的音频都有标签注明0或1,0表示没有鸟叫,1表示有鸟叫。需要根据开发集去预测评估集的标签,鼓励预测0至1之间的概率值,该课题的评估方法使用的是AUC。 这几个数据集之间使用的录音设备、环境状况、样本个数、正负样本比例等都不同,本赛题更想要的是泛化性能

2. 音频数据

给定的三个开发集和三个评估集,每个数据集分别属于同一个音频检测项目。每个数据集中的每个音频长度都是10s,采样率都是44.1k,单声道,并分别标注为0或1,表明无鸟叫或有鸟叫。

值得注意,人工标注的可能会有一定的错误,一般可以认为96.7%是没问题的,对于BirdVox数据集可以认为99.5%或更好。

3. 开发集

3.1 Field recordings, worldwide (“freefield1010”) - a collection of 7,690 excerpts from field recordings around the world, gathered by the FreeSound project, and then standardised for research. This collection is very diverse in location and environment, and for the BAD Challenge we have annotated it for the presence/absence of birds.

Download: [data labels] • [audio files (5.8 Gb zip)] (or [via bittorrent])

3.2 Crowdsourced dataset, UK (“warblrb10k”) - 8,000 smartphone audio recordings from around the UK, crowdsourced by users of Warblr the bird recognition app. The audio covers a wide distribution of UK locations and environments, and includes weather noise, traffic noise, human speech and even human bird imitations.

Download: [data labels] • [audio files (4.3 Gb zip)] (or [via bittorrent])

3.3 Remote monitoring flight calls, USA (“BirdVox-DCASE-20k”) - 20,000 audio clips collected from remote monitoring units placed near Ithaca, NY, USA during the autumn of 2015, by the BirdVox project. More info about BirdVox-DCASE-20k

Download: [data labels] • [audio files (15.4 Gb zip)]

4. 评估集

4.1 Crowdsourced dataset, UK (“warblrb10k”) - a held-out set of 2,000 recordings from the same conditions as the Warblr development dataset.

Download: audio files (1.3 GB zip)

4.2 Remote monitoring dataset, Chernobyl (“Chernobyl”) - 6,620 audio clips collected from unattended remote monitoring equipment in the Chernobyl Exclusion Zone (CEZ). This data was collected as part of the TREE (Transfer-Exposure-Effects) research project into the long-term effects of the Chernobyl accident on local ecology. The audio covers a range of birds and includes weather, large mammal and insect noise sampled across various CEZ environments, including abandoned village, grassland and forest areas.

Download: audio files (5.3 GB zip)

4.3 Remote monitoring night-flight calls, Poland (“PolandNFC”) - 4,000 recordings from Hanna Pamuɫa’s PhD project of monitoring autumn nocturnal bird migration. The recordings were collected every night, from September to November 2016 on the Baltic Sea coast, Poland, using Song Meter SM2 units with microphones mounted on 3–5 m poles. For this challenge, we use a subset derived from 15 nights with different weather conditions and background noise including wind, rain, sea noise, insect calls, human voice and deer calls.

Download: audio files (2.3 Gb zip)

5. baseline

本次官方提供的Baselline是基于去年的比赛Bird Aduio Detection 2017[6]的第一名的代码,BAD 2018 baseline代码的地址是[4]

6. 提交

本次提交预测结果的地址是[5]。评估集大约1.2w个样本,结果预览只展示1000个样本的结果。

参考资料

  1. http://dcase.community
  2. http://dcase.community/challenge2018/task-bird-audio-detection
  3. https://groups.google.com/forum/#!forum/dcase-discussions
  4. https://github.com/DCASE-REPO/bulbul_bird_detection_dcase2018
  5. http://lsis-argo.lsis.org:8005
  6. http://machine-listening.eecs.qmul.ac.uk/bird-audio-detection-challenge/
  7. http://machine-listening.eecs.qmul.ac.uk/bird-audio-detection-challenge-results/
  8. http://c4dm.eecs.qmul.ac.uk/events/badchallenge_results/
  9. https://jobim.ofai.at/gitlab/gr/bird_audio_detection_challenge_2017/tree/master
  10. http://yucc.me/p/62100baa/
  11. http://yucc.me/p/d61fa53b/
谢谢你!