DCASE 2020 Challenge Task 2 Development Dataset

Description This dataset is the "development dataset" for the DCASE 2020 Challenge Task 2 "Unsupervised Detection of Anomalous Sounds for Machine Condition Monitoring" [task description]. The data comprises parts of ToyADMOS and the MIMII Dataset consisting of the normal/anomalous operating sounds of six types of toy/real machines. Each recording is a single-channel (proximately) 10-sec length audio that includes both a target machine's operating sound and environmental noise. The following six types of toy/real machines are used in this task: Toy-car (ToyADMOS) Toy-conveyor (ToyADMOS) Valve (MIMII Dataset) Pump (MIMII Dataset) Fan (MIMII Dataset) Slide rail (MIMII Dataset) Recording procedure The ToyADMOS consists of normal/anomalous operating sounds of miniature machines (toys) collected with four microphones, and the MIMII dataset consists of those of real-machines collected with eight microphones. Anomalous sounds in these datasets were collected by deliberately damaging target machines. For simplifying the task, we used only the first channel of multi-channel recordings; all recordings are regarded as single-channel recordings of a fixed microphone. The sampling rate of all signals has been downsampled to 16 kHz. From ToyADMOS, we used only IND-type data that contain the operating sounds of the entire operation (i.e., from start to stop) in a recording. We mixed a target machine sound with environmental noise, and only noisy recordings are provided as training/test data. For the details of the recording procedure, please refer to the papers of ToyADMOS and MIMII Dataset. Data We first define two important terms in this task: Machine Type and Machine ID. Machine Type means the kind of machine, which in this task can be one of six: toy-car, toy-conveyor, valve, pump, fan, and slide rail. Machine ID is the identifier of each individual of the same type of machine, which in the training dataset can be of three or four. Each machine ID's dataset consists of (i) around 1,000 samples of normal sounds for training and (ii) 100-200 samples each of normal and anomalous sounds for the test. The given labels for each training/test sample are Machine Type, Machine ID, and condition (normal/anomaly). Machine Type information is given by directory name, and Machine ID and condition information are given by their respective file names. Directory structure When you unzip the downloaded files from Zenodo, you can see the following directory structure. As described in the previous section, Machine Type information is given by directory name, and Machine ID and condition information are given by file name, as: /dev_data /ToyCar /train (Only normal data for all Machine IDs are included.) /normal_id_01_00000000.wav ... /normal_id_01_00000999.wav /normal_id_02_00000000.wav ... /normal_id_04_00000999.wav /test (Normal and anomaly data for all Machine IDs are included.) /normal_id_01_00000000.wav ... /normal_id_01_00000349.wav /anomaly_id_01_00000000.wav ... /anomaly_id_01_00000263.wav /normal_id_02_00000000.wav ... /anomaly_id_04_00000264.wav /ToyConveyor (The other Machine Types have the same directory structure as ToyCar.) /fan /pump /slider /valve The paths of audio files are: "/dev_data/<Machine_Type>/train/normal_id_<Machine_ID>_[0-9]+.wav" "/dev_data/<Machine_Type>/test/normal_id_<Machine_ID>_[0-9]+.wav" "/dev_data/<Machine_Type>/test/anomaly_id_<Machine_ID>_[0-9]+.wav" For example, the Machine Type and Machine ID of "/ToyCar/train/normal_id_01_00000000.wav" are "ToyCar" and "01", respectively, and its condition is normal. The Machine Type and Machine ID of "/fan/test/anomaly_id_00_00000000.wav" are "fan" and "00", respectively, and its condition is anomalous. Baseline system 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. Conditions of use This dataset was created jointly by NTT Corporation and Hitachi, Ltd. and is available under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license. Publication If you use this dataset, please cite all the following three papers: 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 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," in Proc. 5th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE), 2020. [pdf] Feedback If there is any problem, please contact us: Yuma Koizumi, [e-mail] Yohei Kawaguchi, [e-mail] Keisuke Imoto, [e-mail]

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」的开发集（development dataset）[任务说明]。该数据集由ToyADMOS数据集（ToyADMOS）与MIMII数据集（MIMII Dataset）的部分内容组成，涵盖六类玩具/实体机器的正常与异常运行声音。每条录音为近似单通道、时长10秒的音频，同时包含目标机器的运行声与环境噪声。本次任务使用的六类机器如下：玩具车（ToyADMOS数据集）、玩具传送带（ToyADMOS数据集）、阀门（MIMII数据集）、泵（MIMII数据集）、风扇（MIMII数据集）、滑轨（MIMII数据集）。

### 录制流程
ToyADMOS数据集通过4个麦克风采集微型玩具机器的正常与异常运行声音，MIMII数据集则通过8个麦克风采集实体工业机器的对应声音。上述数据集的异常声音均通过人为损坏目标机器的方式采集得到。为简化任务设置，本次仅使用多通道录音的第一声道，所有录音均视为固定麦克风采集的单通道音频。所有信号的采样率均下采样至16 kHz。从ToyADMOS数据集中，本次仅选取包含完整运行周期（即从启动到停机全过程）声音的IND类型数据。我们将目标机器声音与环境噪声进行混合，仅提供带噪录音作为训练与测试数据。关于录制流程的详细细节，请参阅ToyADMOS与MIMII数据集的相关学术论文。

### 数据说明
本任务首先定义两个核心术语：机器类型（Machine Type）与机器ID（Machine ID）。机器类型指机器的品类，本次任务包含以下六类：玩具车、玩具传送带、阀门、泵、风扇及滑轨。机器ID指同类型机器的个体唯一标识符，训练数据集的机器ID数量为3或4个。每个机器ID对应的数据集包含：(i) 约1000条用于模型训练的正常声音样本，以及(ii) 测试集各包含100~200条正常与异常声音样本。每个训练/测试样本的标注信息包括机器类型、机器ID与运行状态（正常/异常）。其中机器类型信息由目录名提供，机器ID与运行状态信息则由文件名提供。

### 目录结构
当从Zenodo解压下载的数据集文件后，将看到如下目录层级结构。如前文所述，机器类型信息由目录名标识，机器ID与运行状态信息由文件名标识，具体格式如下：

/dev_data
/ToyCar
/train # 仅包含所有机器ID的正常训练数据
normal_id_01_00000000.wav
...
normal_id_01_00000999.wav
normal_id_02_00000000.wav
...
normal_id_04_00000999.wav
/test # 包含所有机器ID的正常与异常测试数据
normal_id_01_00000000.wav
...
normal_id_01_00000349.wav
anomaly_id_01_00000000.wav
...
anomaly_id_01_00000263.wav
normal_id_02_00000000.wav
...
anomaly_id_04_00000264.wav
/ToyConveyor # 其余机器类型的目录结构与ToyCar完全一致
/fan
/pump
/slider
/valve

音频文件的标准路径格式为：
"/dev_data/<Machine_Type>/train/normal_id_<Machine_ID>_[0-9]+.wav"
"/dev_data/<Machine_Type>/test/normal_id_<Machine_ID>_[0-9]+.wav"
"/dev_data/<Machine_Type>/test/anomaly_id_<Machine_ID>_[0-9]+.wav"
举例如：路径"/ToyCar/train/normal_id_01_00000000.wav"对应的机器类型为「ToyCar」、机器ID为「01」，运行状态为正常；路径"/fan/test/anomaly_id_00_00000000.wav"对应的机器类型为「fan」、机器ID为「00」，运行状态为异常。

### 基线系统
GitHub仓库[URL]中提供了简易基线系统。该基线系统为入门级的基础方案，可在本任务2的数据集上获得合理的检测性能，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。

### 使用许可
本数据集由NTT株式会社与日立有限公司联合创建，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）进行公开分发。

### 引用规范
若使用本数据集，请务必引用以下三篇学术论文：
1. 小泉悠真、斋藤翔一郎、原田升、植松久、井本启介. ToyADMOS：面向异常声音检测的微型机器运行声音数据集[C]//2019年音频与声学信号处理应用研讨会论文集. 2019. [pdf]
2. Harsh Purohit、田边亮、市毛健二、远藤隆、二宫悠纪、末房香织、川口洋平. MIMII数据集：面向工业机器故障排查与检测的声音数据集[C]//第4届声学场景检测与分类研讨会论文集. 2019. [pdf]
3. 小泉悠真、川口洋平、井本启介、中村俊树、二宫悠纪、田边亮、Harsh Purohit、末房香织、远藤隆、安田正博、原田升. DCASE2020挑战赛任务2说明与讨论：面向机器状态监测的异常声音无监督检测[C]//第5届声学场景检测与分类研讨会论文集. 2020. [pdf]

### 问题反馈
如有任何疑问或问题，请联系以下人员：
- 小泉悠真，邮箱：[e-mail]
- 川口洋平，邮箱：[e-mail]
- 井本启介，邮箱：[e-mail]

搜集汇总

背景与挑战

背景概述

该数据集是DCASE 2020挑战赛任务2的开发数据集，包含六种玩具/真实机器的正常和异常操作声音，用于无监督的机器状态监测。数据集提供约1000个正常声音样本用于训练，以及100-200个正常和异常声音样本用于测试，采样率为16kHz，单声道，时长约10秒。