Automated TB diagnosis

Developed an automated system to diagnose TB using Xrays and biomarkers


Current tests for TB in Pakistan are inefficient, complicated, and relatively expensive to perform. This often means the disease has often worsened by the time it is detected and additional delays in detecting the exact location of the disease further contribute to this.


TPI developed novel computational predictive algorithms and machine learning based techniques to efficiently detect TB using biomarkers and xrays


A person's biomarkers and Xray image is taken as input

Preprocessing of data is performed (such as the lungs are segmented from the Xray)

Our trained model is used to detect the presence of TB and the area of the lungs infected by it