NPOL has developed a Throat Infection Analyser (ATIA) using acoustic data processing techniques used in defence applications. The ATIA aims to detect upper respiratory tract infections by non-invasive means of acoustic scanning of word(s) spoken by a person.
It has been proven that CoOVID-19 (Sars Cov-2) virus infection develops initially in upper respiratory tract and detection of infection may give early information even before symptoms like fever etc. develops. For day to day office working, this could give non-contact screening of personnel entering an establishment.
Initially using a microphone connected to computer/ mobile etc. voice samples for healthy person are recorded as per SOP and added in data base.
The recorded data is subjected to acoustic analysis techniques. For carrying out the analysis, signal processing techniques such as Spectrum analysis, Tonal analysis, Modulation and Phase analysis, Mel Frequency Cepstral Coefficients (MFCC) analysis are carried out to generate results for a person.
At any point of time, live voice samples of a person can be compared with data base and distortion in voice due to throat infection is easily captured.
The technique can have potential applications in Hospitals/ clinics, Industries/ offices and also overall pandemic management. However initial recording of voice samples is required.
Presently collection of patients Data has been initiated at Kochi, Delhi and Mumbai through DRDO labs and Industry partner. The data is being processed centrally. Further a US based firm TELLING.AI is also known to be working on similar application. Possibilities of cooperation being worked out.
