Region Duplication Detection Based On Statistical Features of Image

Abstract

Abrupt boom in digital world has led to an instant increase in the popularity of digital images. Easy availability of image tampering tools like Picasa, Adobe Photoshop and Gimp etc. have made image tampering widespread. As such detecting tampering in images has become an active area of research. Region duplication is most common image tampering technique because of the ease with which it can be carried out. Available techniques for region duplication detection fail to accurately locate the tampered region and lack robustness. This paper proposes duplicate region detection method based on statistical texture features using gray level co-occurrence matrix (GLCM) and gray level run length matrix (GLRLM) features. The method divides the forged image into overlapping blocks, calculate texture features based on GLCM and GLRLM of each block. Feature vectors thus obtained for each block are lexicographically sorted. Blocks with similar features are identified using feature distances. Post processing isolates the duplicate regions. Experimental results establish that the proposed method using GLRLM features can precisely locate duplicate regions in image and can effectively withstand the common post processing operation like jpeg compression, blurring, brightness and contrast change with reduced computation complexity.