Glaucoma, a serious vision illness that can result in irreversible vision loss, has a huge influence on the health of people all over the world because it causes damage to the optic nerve. The purpose of this research is to improve the accuracy and efficiency of screenings and lessen the impact of the condition on a global scale. Convolutional Neural Networks (CNNs) are utilized in this investigation to improve early glaucoma identification. Several different CNN architectures and data pre-processing methods were investigated using the PAPILA dataset. These methods and architectures included VGG19, VGG16, InceptionV3, Xception, EfficientNetB0, ResNet50, and DenseNet121. It is recommended to use VGG16 for generic detection that does not require any specific pre-processing. According to the study’s findings, EfficientNetB0 attained a remarkable area under the curve (AUC) of 78.29% (0.78), highlighting its potential utility in increasing glaucoma detection.