Authors: Haroon Mahmood; Manal Iftikhar; Aamir Wali; Arshad Ali; Maryam Gulzar
This study explores using artificial intelligence to diagnose tuberculosis (TB) by analyzing cough sounds. The researchers developed a system that first distinguishes between coughs and background noise, then classifies the cough as either TB-related or not. They collected cough recordings from patients at a hospital in Lahore, Pakistan, and tested the system using various machine learning models. The results showed that including patient information like demographics and clinical data improved the accuracy of the diagnosis, achieving up to 97% accuracy with one model. This approach could help community health workers detect TB more effectively and affordably.
Download Resource
Resource Topic: Technology, Tuberculosis, mHealth and Technology
Resource Type: Qualitative
Year: 2024
Region: Asia
Country: Pakistan
Publisher May Restrict Access: No