Capstone Data Science Project
Budget: £250 – £750 GBP
Aims: Implementing several classification-based conformal predictors to detect coughing and sneezing sounds, and comparing their performance validity and efficiency on the recently released ESC dataset.
Background:
Non-speech body sounds recognition had not received major attentions, despite containing valuable information about the person's health conditions.
This project aims at detecting the two most popular infectious events, coughs and sneezes, from raw recordings.
The challenge for such objective is the variety of coughs and sneezes, amongst other non-speech body sounds. Hence, the confidence measure of conformal prediction would be useful in such scenario.
Background:
Non-speech body sounds recognition had not received major attentions, despite containing valuable information about the person's health conditions.
This project aims at detecting the two most popular infectious events, coughs and sneezes, from raw recordings.
The challenge for such objective is the variety of coughs and sneezes, amongst other non-speech body sounds. Hence, the confidence measure of conformal prediction would be useful in such scenario.