Image Classification and Detection of Artificial Images Using CNN Models

In the world of digital media, the introduction of AI –generated generated artificial images has created serious problems in discriminating between actual and constructed visual information. The ease with which anyone may utilize these innovations for creating propaganda might spread fear and anarchy as a result of their rapid growth. Therefore, in this era of social media, having a reliable mechanism to distinguish between authentic and fraudulent content is essential. These photos, which often appear identical from legitimate ones, represent a danger to the integrity of digital media, potentially leading to deception and crime. Due to this the approaches like as DL are being employed more frequently to discriminate between real and fake faces, yielding more accurate and consistent outcomes. This study investiagtes and examines the perfomance of CNN architectures like AlexNet, VGG19, EfficientNet, ResNet50 and InceptionV3 for the classification of real and fake images. The performance is assessed using the accruacy,precision and F1 Score using the ArtiFact dataset, which is available publicly. The ResNet50 outperforms the other models with the highest classification accuracy.