SNRCN2: Steganalysis noise residuals based CNN for source social network identification of digital images

Kapil Rana, Gurinder Singh, Puneet Goyal

Abstract: Highlights•A novel CNN method SNRCN2 for source social media network identification of an image.•Steganalysis-based filters used to suppress image content and extract noise residuals.•The purpose of extracting noise residual is to highlight the social network artifacts.•The noise residuals are passed to CNN for high-level feature extraction.•Results show that the proposed network consistently outperforms the existing techniques.