Texture Operator based Digital Image Splicing Detection using ELTP Technique

Abstract

Social networking sites play a significant role in news viral. A lot of information is fake. Image editing tools play an essential role in the easy availability of these tools and software. So image authentication is the need of the time. In this paper, we proposed a texture-based image splicing detection, ELTP (Enhanced local ternary pattern) used for feature extraction, which gives us a better response than LBP. The local ternary pattern used ternary value (-1, 0, 1) code for assessment, advanced to LBP binary value code. ELTP is more sensitive compare to noise. In our proposed technique, first, we convert the RGB image into YCbCr, and the chrominance component used for feature extraction. In feature extraction, first, we implement a standard deviation filter to highlight the abrupt change in the image and ELTP to extract the features. SVM is used as a classifier gives better to some of the states of the art method.