Forensic analysis of ordered data structures on the example of JPEG files

Abstract

JPEG file format standards define only a limited number of mandatory data structures and leave room for interpretation. Differences between implementations employed in digital cameras, image processing software, and software to edit metadata provide valuable clues for basic authentication of digital images. We show that there exists a realistic chance to fool state-of-the-art image file forensic methods using available software tools and introduce the analysis of ordered data structures on the example of JPEG file formats and the EXIF metadata format as countermeasure. The proposed analysis approach enables basic investigations of image authenticity and documents a much better trustworthiness of EXIF metadata than commonly accepted. Manipulations created with the renowned metadata editor ExifTool and various image processing software can be reliably detected. Analysing the sequence of elements in complex data structures is not limited to JPEG files and might be a general principle applicable to different multimedia formats.