Source Camera Identification Using Interninsic Fingerprints

The identification of an image acquisition device becomes a very important in digital image forensics. In this paper main focus is on identification of digital camera and provides a new approach of identification. As digital camera uses many components in image generation process such as lens, sensors, CFA (colour filter array) e.t.c. Each component has its unique characteristics that can be known as fingerprints. By extracting that fingerprint identification of digital camera can be done. In this paper identification is done on the basis of sensor noise and CFA unique characteristics. From sensor photo-response non-uniformity (PRNU) is considered because it was introduced by Luk´aˇs et al. [2] for identification of digital camera. So in this work PRNU (Photo Response Non Uniformity) and 12 features from CFA are used. Classification is done by using SVM classifier (Support Vector Machine).  Experimental analysis shows that the proposed method has good potential for identification of digital camera.