Classification of Lung Chest X-Ray Images Using Deep Learning with Efficient Optimizers

This paper investigates the application of several deep learning architectures such as VGG-16, VGG-19, ResNet50, and Xception Net for lung chest X-ray images, with 5, 10, and 15 epochs, and different optimizers such as Adam, SGD, and RMSProp, and a learning rate of 0.0001. The investigation finds that when adaptive gradient is used, the VCG-16 architecture achieves 68% accuracy; VCG-19 achieves 67% accuracy; ResNet50 achieves 98.67% accuracy; and the Xception Net architecture achieves less than 50% accuracy. With further experimentation using 5, 10, and 15 epochs and optimizers such as Adam, SGD, and RMSProp, a 100% accuracy was achieved with 15 epochs for the VGG-16, VGG-19, and ResNet-50 architectures. However, Xception Net has been able to achieve only 70% accuracy with these optimizers.