Estimation of Spliced Images in Photographs

Photographs are considered to be a trustful document which is used to store life events. But the recent advance in image editing softwares causes image manipulation more easier and common since it will not leave any visual clue of having it tampered. So it become difficult to identify whether an image is authentic or not. There is mainly three types of image forgery, they are image retouching, copy move and image splicing. There are different methods available to check the authenticity of the image. In this paper we propose a method to detect splicing in photo consisting of human. First we detect all the faces present in the image then it is transformed into different color space. Then illuminant map of each color space is obtained. Extract different features such as texture color and some of the image quality measures(IQM). Naive Bayes classifier is used to detect whether the image is real or fake.