SOUNDCAM
收藏资源简介:
SOUNDCAM数据集由斯坦福大学创建,包含5000个真实的房间脉冲响应和2000个音乐录音,用于研究室内声学特性。数据集涵盖了三种不同类型的房间:受控声学实验室、野外客厅和会议室,每个房间中都有不同位置的人。该数据集可用于检测和识别人类,以及跟踪他们的位置,适用于虚拟/增强现实和智能家居助手等领域。数据集的创建过程涉及在不同房间中记录人的位置和声学响应,以及使用Azure Kinect DK RGBD相机捕捉人的姿态数据。SOUNDCAM数据集旨在解决通过声学信号进行人类定位、识别和检测的问题,为声学研究提供了宝贵的资源。
The SOUNDCAM dataset was developed by Stanford University, which contains 5,000 real room impulse responses and 2,000 musical recordings for research into indoor acoustic characteristics. The dataset covers three distinct types of rooms: controlled acoustic laboratories, real-world living rooms and conference rooms, with people located at different positions within each room. This dataset can be utilized for human detection, recognition and position tracking, and is applicable to domains such as virtual/augmented reality and smart home assistants. The creation of the SOUNDCAM dataset involved recording the positions of human subjects and their corresponding acoustic responses across different rooms, as well as capturing human pose data using the Azure Kinect DK RGBD camera. The SOUNDCAM dataset is designed to address the problems of human localization, recognition and detection via acoustic signals, providing a valuable resource for acoustic research.




