Fake Image Detection in Fake News Using Convolutional Neural Network (CNN)

’Fake news’ refers to false, inaccurate, or misleading information that spreads as real news. Fake news primarily aims to affect societies and individuals by spreading false or misleading information. Fake news comes in text, images, videos, and audio formats. Detecting fake news is a research challenge that requires a promising solution to assure the accuracy of the information. Recently, promising results have been generated using artificial intelligence to detect fake news in multimedia. Computer vision, natural language processing techniques, audio processing, and machine learning algorithms are all techniques used in artificial intelligence to help accurately differentiate accurate news from fake news. However, The paper proposed a Convolutional Neural Networks (CNNs) model that can detect fake images in the context of fake news. The paper also utilized various preprocessing techniques, including error level analysis (ELA), gradient, and noise analysis. These techniques are compared to discern their effectiveness in enhancing the model’s performance. The result indicates that the proposed CNN model achieved an overall accuracy of 91.38% on the CASIA V2.0 dataset with error level analysis (ELA) as the preprocessing technique. The results indicate that ELA is an effective preprocessing technique for extracting features that can be used to identify fake images.