Artificial Intelligence based Deep Architecture for Tuberculosis Detection

---

Abstract

Artificial Intelligence-based system for Tuberculosis (TB) detection has been proposed in this work. TB detection in the early stage can help in mitigating the effects in organ damage of kidneys, liver, spine, and brain, thus reducing human deaths. Manual diagnostic through radiologists can be misdiagnosis due to human error. Designing an artificial intelligence-based decision support system can help in the accurate prediction of TB through lung Chest X-ray (CXR) image. In the proposed work, five different kinds of image enhancement techniques were applied to the publicly available image dataset. The dataset is balanced using a data balancing approach with eight different kinds of data augmentation applied. Six different modified network architectures of Convolutional Neural Network(CNN) is utilized. Comparative Studies of five image enhancement techniques in behaviour response with CNN architectures are evaluated for best performance. The obtained result showed Gamma corrected image with Modified ChexNet performed the best with 96.3% accuracy, 95.3%precision and 97% recall.

---

Keywords

P AI, TB, CXR, CNN, accuracy, recall, precision