DCASE 2018, Task 5: Monitoring of domestic activities based on multi-channel acoustics - Development dataset

This repository contains the development data of task 5 of the DCASE 2018 challenge. The dataset is a derivative of the SINS database. The SINS database contains a continuous recording of one person living in a vacation home over a period of one week. The recordings were manually annotated on daily activity level: "Cooking", "Dishwashing", "Eating", "Social activity (visit, phone call)", "Vacuum cleaning", "Watching TV", "Working", "Presence" and "Absence". More information can be found on (please cite this papers when using the dataset): G. Dekkers, S. Lauwereins, B. Thoen, M. W. Adhana, H. Brouckxon, T. van Waterschoot, B. Vanrumste, M. Verhelst, and P. Karsmakers, “The SINS database for detection of daily activities in a home environment using an acoustic sensor network,” in Proceedings of the Detection and Classification of Acoustic Scenes and Events 2017 Workshop (DCASE2017), Munich, Germany, November 2017, pp. 32–36. G. Dekkers, L. Vuegen, T. van Waterschoot, B. Vanrumste, and P. Karsmakers, “DCASE 2018 Challenge - Task 5: Monitoring of domestic activities based on multi-channel acoustics,” KU Leuven, Tech. Rep., July 2018. The derivative of the SINS database, 'DCASE 2018 – Task 5 development dataset' consists of data collected by 4 microphone arrays in the combined living room and kitchen area. The continuous recordings were split into audio segments of 10s. These audio segments are provided as individual files along with the ground truth. In total 72984 segments are made available, leading to approximately 200 hours of data. More information about the challenge and the specific dataset can be found here. Information solely related to the content of the dataset is available in 'DCASE2018-task5-dev.doc.zip'. By accessing or using this database, the user accepts the provided EULA (available in DCASE2018-task5-dev.doc.zip).本仓库包含声学场景检测与分类挑战赛2018（Detection and Classification of Acoustic Scenes and Events 2018，简称DCASE 2018）任务5的开发数据集。该数据集衍生自SINS数据库（SINS database）。SINS数据库记录了一名受试者在度假住宅内为期一周的连续声学采集数据，采集所得数据已按日常活动层级完成人工标注，标注类别包括：烹饪（Cooking）、洗碗（Dishwashing）、用餐（Eating）、社交活动（Social activity (visit, phone call)，涵盖访客拜访与电话通话）、吸尘（Vacuum cleaning）、看电视（Watching TV）、工作（Working）、在场（Presence）与不在场（Absence）。使用该数据集时请引用以下两篇文献：1. G. Dekkers、S. Lauwereins、B. Thoen、M. W. Adhana、H. Brouckxon、T. van Waterschoot、B. Vanrumste、M. Verhelst及P. Karsmakers，《基于声学传感器网络的家庭环境日常活动检测SINS数据库》，收录于2017年声学场景检测与分类研讨会（DCASE2017）论文集，德国慕尼黑，2017年11月，第32-36页。2. G. Dekkers、L. Vuegen、T. van Waterschoot、B. Vanrumste及P. Karsmakers，《DCASE 2018挑战赛任务5：基于多通道声学的家庭活动监测》，KU Leuven，技术报告，2018年7月。本SINS数据库的衍生数据集——“DCASE 2018任务5开发数据集”，由部署于客厅与厨房复合区域的4组麦克风阵列采集所得。原始连续录音被切分为时长10秒的音频片段，以独立文件形式发布，并附带对应的基准真值（ground truth）。本次共发布72984条音频片段，总时长约200小时。关于挑战赛与该数据集的更多详情可查阅指定链接。仅与数据集内容相关的完整说明可在“DCASE2018-task5-dev.doc.zip”压缩包中获取。用户访问或使用本数据库即视为同意其中附带的最终用户许可协议（End User License Agreement，简称EULA），该协议可于该压缩包内获取。