Development of the Error Level Analysis Forensic Tool for Images Shared Over Messaging and Social Networking Applications
Error level analysis is considered as a forensic tool that reveals forged regions in JPEG images. Though being
regarded as a simple method; it is considered as an indispensable block in digital image examination routine. Conversely, it
has poor performance over images that have been heavily compressed such as images shared over messaging and social
networking applications. In this paper, we develop the error level analysis method in order to reveal forgeries in these
specific low-quality images. The error-level map is generated and obtained by analyzing the frequency coefficients of the
image using the discrete cosine transform. AC coefficients are limited in order to minimize high frequency noise. A specific
database of images captured by mobile phones, altered and then shared over well-known messaging and social networking
applications is created. The developed method is then implemented and examined over the database. Obtained results show
higher readability and better image modifications discrimination than those obtained with the original error analysis method.
Keywords - Error Level Analysis; Digital Image Processing; Discrete Cosine Transform.