The Second DiCOVA Challenge: Dataset and performance analysis for\n COVID-19 diagnosis using acoustics

Neeraj Kumar Sharma, Srikanth Raj Chetupalli, Debarpan Bhattacharya, Debottam Dutta, Pravin Mote, Sriram Ganapathy

arXiv (Cornell University) · 2021 · 10 citations · 0 references

DOIFull text

Open access

Concepts

Abstract

The Second Diagnosis of COVID-19 using Acoustics (DiCOVA) Challenge aimed at\naccelerating the research in acoustics based detection of COVID-19, a topic at\nthe intersection of acoustics, signal processing, machine learning, and\nhealthcare. This paper presents the details of the challenge, which was an open\ncall for researchers to analyze a dataset of audio recordings consisting of\nbreathing, cough and speech signals. This data was collected from individuals\nwith and without COVID-19 infection, and the task in the challenge was a\ntwo-class classification. The development set audio recordings were collected\nfrom 965 (172 COVID-19 positive) individuals, while the evaluation set\ncontained data from 471 individuals (71 COVID-19 positive). The challenge\nfeatured four tracks, one associated with each sound category of cough, speech\nand breathing, and a fourth fusion track. A baseline system was also released\nto benchmark the participants. In this paper, we present an overview of the\nchallenge, the rationale for the data collection and the baseline system.\nFurther, a performance analysis for the systems submitted by the $16$\nparticipating teams in the leaderboard is also presented.\n