Kajian Penerapan Semi-Automated Evaluation Based on Similarity pada Investigasi Digital Image Forensics

Abstract

Along with the development of information technology, is now widely available for image processing applications are easy to use. It has an impact on the process of manipulation and modification of the image to be more easily so it is quite difficult to determine the authenticity of an image. Image evaluation methods have been developed automatically. The weaknesses are irrelevant parts of the image will be processed so as not efficient in terms of processing time. User Role (semi-automatic) for assessing the relevance of an image section can be helpful in the evaluation method. Similarity-based approach, one of them using the Euclidean distance, either automatic or semi-automatic allows forensic computer analysis process becomes faster. Preparation of the ranking of the data that has similarities allows users to investigate the authenticity of the picture. In this paper tried to assess the implementation of semi-automatic evaluation of similarity based on a forensic image. In the method of analysis, there are two stages, namely the conduct of the evaluation results ranking image block that requires verification of the user so that the second stage is only relevant image blocks are analyzed automatically. The results showed that the use of semi-automatic evaluation methods based on computer forensic similarity not only be done on the data string but also the image data.