Morphological Pattern Spectrum Based Image Manipulation Detection

This paper presents an image histogram-based image manipulation detection method in android. The method consists of using a mathematical morphological-based algorithm to extract pictorial feature information from an original digital image. This will be useful in situations where it is important to find evidence of specific events such as investigation of crimes or simply image comparison. The morphological pattern spectrum was implemented in android platform using Java and OpenCV. Analysis of manipulated images indicated that the proposed detection method was able to clearly identify the differences from the original images. The results show that the proposed technique has sufficient ability to distinguish the very slight manipulation of upto one pixel size.