An Object-Based Detection of Splicing Forgery using Color Illumination Inconsistencies

This digital imaging era has become ingrained in people to share images and videos with their loved ones via social networking sites. On the other hand, photo editing software innovation allows anyone to easily and quickly modify multimedia content. As a result, people have become accustomed to alter digital content for the sake of amusement or to attract the attention of others. As a result, Digital Image Forensics emerges to provide trustworthiness to the content shared via the internet. This paper proposes efficient tampering detection and localization method for detecting splicing attacks using color illumination inconsistencies. First, the image's color illuminant features extract using the grayness index in rg-chromaticity space. The color illuminants of each color channel are then combined with a logistic regression model to reveal tampered objects in the image. The experimental results show that the proposed method outperforms the state-of-the-art.