D N Keerthana , D Jayasree , V Jyothirmayee , C Jyoshna Reddy , S Ankitha 1 2 (April - June) Preventing the spread of tuberculosis TB and enhancing patient outcomes are largely dependent on early detection This presents a new method for detecting tuberculosis by utilizing convolutional neural networks to examine Xray pictures of the chest To extract and learn hierarchical information from xray images a critical component for effective predictions the CNNs architecture has been meticulously built By examining the interpretability of the CNN model the study sheds light on the characteristics and patterns that the network considers important when forming predictions For medical professionals to trust the model and use it as a useful diagnostic tool this feature is essential Firstly it outlines the limitations of conventional TB diagnostic methods emphasizing the need for more accurate efficient and accessible solutions Copyright (c) World Journal of Pharmaceutical Seiences. All rights reserved 2026