DCASE 2020 Challenge Task 2 Evaluation Dataset

<strong>Description</strong> This dataset is the "evaluation dataset" for the <strong>DCASE 2020 Challenge Task 2 "Unsupervised Detection of Anomalous Sounds for Machine Condition Monitoring" </strong>[task description]. In the task, three datasets have been released: "development dataset", "additional training dataset", and "evaluation dataset". This evaluation dataset was the last of the three released. This dataset includes around 400 samples for each Machine Type and Machine ID used in the evaluation dataset, none of which have a condition label (i.e., normal or anomaly). The recording procedure and data format are the same as the development dataset and additional training dataset. The Machine IDs in this dataset are the same as those in the additional training dataset. For more information, please see the pages of the development dataset and the task description. After the DCASE 2020 Challenge, we released the ground truth for this evaluation dataset. <strong>Directory structure</strong> Once you unzip the downloaded files from Zenodo, you can see the following directory structure. Machine Type information is given by directory name, and Machine ID and condition information are given by file name, as: /eval_data /ToyCar /test (Normal and anomaly data for all Machine IDs are included, but they do not have a condition label.) /id_05_00000000.wav ... /id_05_00000514.wav /id_06_00000000.wav ... /id_07_00000514.wav /ToyConveyor (The other Machine Types have the same directory structure as ToyCar.) /fan /pump /slider /valve The paths of audio files are: "/eval_data/&lt;Machine_Type&gt;/test/id_&lt;Machine_ID&gt;_[0-9]+.wav" For example, the Machine Type and Machine ID of "/ToyCar/test/id_05_00000000.wav" are "ToyCar" and "05", respectively. Unlike the development dataset and additional training dataset, its condition label is hidden. <strong>Baseline system</strong> A simple baseline system is available on the Github repository [URL]. The baseline system provides a simple entry-level approach that gives a reasonable performance in the dataset of Task 2. It is a good starting point, especially for entry-level researchers who want to get familiar with the anomalous-sound-detection task. <strong>Conditions of use</strong> This dataset was created jointly by <strong>NTT Corporation</strong> and <strong>Hitachi, Ltd.</strong> and is available under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license. <strong>Publication</strong> If you use this dataset, please cite <strong>all the following three papers</strong>: Yuma Koizumi, Shoichiro Saito, Noboru Harada, Hisashi Uematsu, and Keisuke Imoto, "ToyADMOS: A Dataset of Miniature-Machine Operating Sounds for Anomalous Sound Detection," in Proc. of IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA), 2019. [pdf] Harsh Purohit, Ryo Tanabe, Kenji Ichige, Takashi Endo, Yuki Nikaido, Kaori Suefusa, and Yohei Kawaguchi, “MIMII Dataset: Sound Dataset for Malfunctioning Industrial Machine Investigation and Inspection,” in Proc. 4th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE), 2019. [pdf] Yuma Koizumi, Yohei Kawaguchi, Keisuke Imoto, Toshiki Nakamura, Yuki Nikaido, Ryo Tanabe, Harsh Purohit, Kaori Suefusa, Takashi Endo, Masahiro Yasuda, and Noboru Harada, "Description and Discussion on DCASE2020 Challenge Task2: Unsupervised Anomalous Sound Detection for Machine Condition Monitoring<em>,"</em> in Proc. 5th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE), 2020. [pdf] <br> <strong>Feedback</strong> If there is any problem, please contact us: Yuma Koizumi, [e-mail] Yohei Kawaguchi, [e-mail] Keisuke Imoto, [e-mail]

**数据集说明** 本数据集为**DCASE 2020挑战赛任务2：面向机器状态监测的异常声音无监督检测**的**评估数据集（evaluation dataset）**[任务说明]。本次任务共发布三类数据集：「开发数据集（development dataset）」、「额外训练数据集（additional training dataset）」与该评估数据集，本评估数据集为三者中最后发布的版本。本数据集针对每种机器类型（Machine Type）与机器ID（Machine ID）均提供了约400条样本，所有样本均未附带状态标签（即正常或异常）。其录制流程与数据格式均与开发数据集、额外训练数据集保持一致，且所含机器ID与额外训练数据集的机器ID完全相同。更多信息请参阅开发数据集与任务说明页面。DCASE 2020挑战赛结束后，我们公开了本评估数据集的真实标签（ground truth）。
**目录结构** 从Zenodo下载并解压文件后，即可看到如下目录结构。机器类型信息由目录名称给出，机器ID与状态信息由文件名标识，具体结构如下：
/eval_data
/ToyCar
/test （包含所有机器ID的正常与异常数据，但未附带状态标签）
/id_05_00000000.wav ...
/id_05_00000514.wav
/id_06_00000000.wav ...
/id_07_00000514.wav
/ToyConveyor （其余机器类型的目录结构与ToyCar一致）
/fan
/pump
/slider
/valve
音频文件的路径格式为：`/eval_data/<Machine_Type>/test/id_<Machine_ID>_[0-9]+.wav`
例如，文件路径`/ToyCar/test/id_05_00000000.wav`对应的机器类型与机器ID分别为「ToyCar」与「05」。与开发数据集和额外训练数据集不同，本数据集隐藏了状态标签。
**基线系统（baseline system）** GitHub仓库[URL]中提供了一套简易基线系统。该基线系统提供了入门级的简易解决方案，可在本任务2的数据集上取得合理的性能表现，尤其适合希望熟悉异常声音检测任务的入门研究者作为起步参考。
**使用条款** 本数据集由NTT公司（NTT Corporation）与日立有限公司（Hitachi, Ltd.）联合创建，采用知识共享署名-非商业性使用-相同方式共享4.0国际许可协议（Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International，CC BY-NC-SA 4.0）进行授权。
**引用要求** 若使用本数据集，请引用以下全部三篇论文：
1. 小泉悠马（Yuma Koizumi）、斋藤翔一郎（Shoichiro Saito）、原田升（Noboru Harada）、植松久（Hisashi Uematsu）、井本圭介（Keisuke Imoto），《ToyADMOS：用于异常声音检测的微型机器运行声音数据集》，发表于2019年IEEE音频与声学信号处理应用研讨会（IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA）。[pdf]
2. 哈什·普魯希特（Harsh Purohit）、田边辽（Ryo Tanabe）、市毛健二（Kenji Ichige）、远藤隆（Takashi Endo）、二木悠希（Yuki Nikaido）、末保香织（Kaori Suefusa）、川口洋平（Yohei Kawaguchi），《MIMII数据集：用于工业机器故障检测与排查的声音数据集》，发表于第4届声学场景检测与分类研讨会（4th Workshop on Detection and Classification of Acoustic Scenes and Events, DCASE），2019年。[pdf]
3. 小泉悠马（Yuma Koizumi）、川口洋平（Yohei Kawaguchi）、井本圭介（Keisuke Imoto）、中村俊树（Toshiki Nakamura）、二木悠希（Yuki Nikaido）、田边辽（Ryo Tanabe）、哈什·普魯希特（Harsh Purohit）、末保香织（Kaori Suefusa）、远藤隆（Takashi Endo）、安田正博（Masahiro Yasuda）、原田升（Noboru Harada），《DCASE2020挑战赛任务2说明与探讨：面向机器状态监测的异常声音无监督检测》，发表于第5届声学场景检测与分类研讨会（5th Workshop on Detection and Classification of Acoustic Scenes and Events, DCASE），2020年。[pdf]
**反馈与联系** 若有任何问题，请联系以下人员：
Yuma Koizumi，邮箱：[e-mail]
Yohei Kawaguchi，邮箱：[e-mail]
Keisuke Imoto，邮箱：[e-mail]