ANALYSIS AND IMPLEMETATION OF SOURCE CAMERA IDENTIFICATION TECHNIQUE

Abstract

An intriguing issue in computerized legal sciences is that given a digital picture, would it be possible to distinguish the camera model which was utilized to get the picture particularly in this age of fraud. In this thesis work we take a look on the technique which can be used to solve this purpose in simpler way. We studied the four diverse procedure and look at their predominance. Simplified techniques are to check for the features which are unique to camera. These features can be utilized by a classifier to distinguish the source camera of a picture. Also, the recognizable feature components were further classified for changes, by analyses, to improve the accuracy rate of the distinguishing proof. This exploration additionally looked at cameras of changed brand, as well as those of the same maker with comparative models. An inherent component of a camera is PRNU (photograph response non-uniformity). Such unique features can be assessed from various pictures taken by the same camera by method for a denoising filtering operation. With the utilization of a color filter array for catching one and only of the three colors channel for every pixel presents shading interpolation error, while the current techniques for extracting PRNU, incorporate the shading interjection mistake as a feature of the PRNU, which leaves opportunity to get better. So we propose another method for extracting PRNU, called Color-Decoupled PRNU (CD-PRNU), by using the distinction between the physical and artificial color channel of the photographs taken by digital cameras that utilization a color filter array for introducing manufactured color channel from the physical ones. Past proposed camera methods calculations are set up to group the sources for a predefined compression factor quality. Conversely, the compression characteristics of the picture v might be changed amid further processing and altering. Local binary pattern applies on the edges extracted from the information picture. The component demonstrated great strength against JPEG pressure quality varieties. The strategy uncovered its capacities in both high and low quality picture.