TBNet: A Reliable and Efficient Neural Network for Tuberculosis Detection

The next generation of MobileNets is presented in this article, which is built on merging various fundamental blocks and an entirely novel architecture design for computer vision and embedded applications. This design increases Tuberculosis (TB) detection while utilising little computing resources. This architecture is compared to other state-of-theart Novel architectures, demonstrating its capability. This neural network builds on the original MobileNet architecture by including Bi-Directional block to extract more features from images for improved detection and outcomes. MobileNet design is the foundation for our architecture; it is built on MobileNet, but the hyperparameters are tailored to the application's needs to identify tuberculosis in human lungs.