Abstract
The general objective of this Thematic Project is the development of theoretical, technical and technological knowledge about biomarkers of respiratory conditions, whose application is practical, cheap and viable for the Brazilian population. Using machine learning techniques, artificial intelligence and audio analysis, we will investigate audio biomarkers for respiratory conditions such as respiratory failure, effects of smoking, and severe asthma. The project aims to develop, in a first phase, mobile computing systems that carry out the task of audio collection and whose data will be subject to audio analysis processes and neural network training; in a second phase, a system for applying the information resulting from learning will be used in a clinical environment, in a process of clinical validation of the tool. There are several research challenges inherent in these goals. In the area of artificial intelligence, we will have to develop systems for classifying, estimating and predicting events; for example, we intend to classify voices with respiratory failure, estimate the level of exhaled carbon monoxide (COex) by smokers and predict the probability of occurrence of asthma exacerbation. In the area of software engineering, we have the challenge of developing a reference architecture for healthcare audio systems. in the area of voice and speech studies, we will need to find discriminating patterns of biomarker conditions, for example, conditions identifying respiratory failure. In health areas, we face the challenges of obtaining audio recordings from patients during follow-up and of integrating clinical work with artificial intelligence techniques, in order to enable the practical application of biomarkers. In order to be able to face these challenges, the project is organized in a matrix form, with three lines of research aimed at developing audio biomarkers by mobile computing for the conditions: (1) respiratory failure; (2) effects of smoking; and (3) severe asthma. These activities are dealt with by four cross-cutting axes of actions in the areas of: (a) data collection; (b) software engineering; (c) acoustic analysis; and (d) machine learning.This Project is based on the research results achieved in the SPIRA project (FAPESP, 2020/06443-5) which, motivated by the COVID-19 pandemic, demonstrated the feasibility of detecting respiratory failure by analyzing audio captured on mobile devices in patients of the first wave of the pandemic. The multidisciplinary research group has the same core of researchers, expanded to deal with questions of the effects of smoking and severe asthma. (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |