Detection And Classification Of Various Image Operations Using Deep Learning Technology

As one of the main medium for information transmission, the digital imagecan be easily tampered during transmission. It is becoming more and more important to identifywhether the given image is an original image or a processed image. In this paper, we propose acompact universal feature based on spatial domain in virtue of some latest image forensic methods and design a multi-class classification scheme using the deep learning technique, to identify and furthermore classify the various normal image operations. According to the experimental results, the proposed method can well detect and classify the multiple common image post-processing operations. And the comparison with the existing feature shows the better performance of the proposed method.