Fiber-optic ESC-50, Gunshot-Firework Event Classification Dataset
收藏资源简介:
Fiber-optic ESC-50数据集由NEC美国实验室公司和纽约州立大学布法罗分校创建,包含2000个样本,分为50个类别,主要用于研究光纤声学识别任务。该数据集通过分布式声学传感技术记录,涵盖了多种环境声音。枪声-烟花事件分类数据集则包含8种真实生活中的声音事件,如枪声、烟花等,用于评估模型在复杂户外环境中的表现。这些数据集通过实验室和真实环境中的光纤传感器记录,旨在解决光纤声学识别中的领域适应和少样本学习问题,广泛应用于环境感知和安全监控领域。
The Fiber-optic ESC-50 dataset was developed by NEC Laboratories America, Inc. and the University at Buffalo, The State University of New York. It contains 2000 samples divided into 50 categories, and is mainly used for research on fiber-optic acoustic recognition tasks. This dataset was recorded via distributed acoustic sensing technology, covering a variety of environmental sounds. The Gunshot-Firework Event Classification dataset includes 8 types of real-life sound events such as gunshots and fireworks, and is utilized to evaluate the performance of models in complex outdoor environments. Recorded using fiber-optic sensors in both laboratory and real-world scenarios, these datasets are designed to tackle domain adaptation and few-shot learning problems in fiber-optic acoustic recognition, and have been widely applied in the fields of environmental perception and security monitoring.




