Double Compression Detection of HEIF Images Using Coding Ghosts

Extensive research on double-compressed image analysis has been performed in image forensics, referring to the widely adopted JPEG coding. However, as HEIF gains popularity for its efficiency in reducing file sizes while maintaining image quality, the lack of methods for detecting double HEIF compression becomes apparent. Traditional JPEG-based techniques are not directly applicable to HEIF due to their distinct coding algorithms. In this study, we build upon Farid’s work on coding ghosts in JPEG images and introduce a method to detect double-aligned compression and estimate initial quantization coefficients in HEIF images. Our experiments show that this method performs effectively when the difference between the first and second quantization parameters (QP) exceeds 5.