DiCOVA Challenge Dataset

DiCOVA Challenge Dataset是由印度科学研究所和P. D. Hinduja医院创建的，旨在通过声音分析加速COVID-19的诊断研究。该数据集包含从COVID-19感染者和非感染者收集的呼吸声音记录，通过网络应用程序在全球范围内收集。数据集分为两个轨道，Track-1专注于咳嗽声音，Track-2则包括呼吸、持续元音发音和数字计数语音记录。数据集的创建过程涉及使用Coswara数据集，并通过网络应用程序收集声音样本。该数据集的应用领域主要集中在通过声音分析辅助COVID-19的诊断，提供一个快速、易于使用且成本有效的诊断工具。

The DiCOVA Challenge Dataset was developed by the Indian Institute of Science and P. D. Hinduja Hospital, with the objective of accelerating COVID-19 diagnostic research via audio analysis. This dataset includes respiratory audio recordings collected from both COVID-19-positive and COVID-19-negative individuals, which were gathered globally through a web application. The dataset is divided into two tracks: Track-1 focuses on cough audio, while Track-2 covers respiratory sounds, sustained vowel pronunciations, and speech recordings of digit counting. The development process leveraged the Coswara Dataset and collected audio samples via the aforementioned web application. The primary application of this dataset is to assist COVID-19 diagnosis through audio analysis, providing a fast, user-friendly and cost-effective diagnostic tool.