Camera Model Identification using Convolutional Neural Network

Abstract

Source camera identification is the procedure of finding out which camera model has been used to capture an image.In the past few years, there has been a high-speed growth of research interest in the field of forensics.In our present work, we have proposed a Deep Learning approach for identifying the camera model of ten cameras as a part of the Camera Model Identification Challenge organized by the Kaggle. com. Through this paper, we have presented a camera model identification method based on Convolutional Neural Network (CNN). In contrast to traditional methods, CNNs can spontaneously and concurrently extricate features and can be trained to classify during the learning process. Source camera identification is used in legal as well as security matters as a proof. As a correlation to previous task, researchers have suggested to utilize the artifacts that are present in the pipeline of camera to gather particular features manually and use them to differentiate between individual devices or camera models. Our task is implemented with 10 camera models from the dataset provided by the IEEE's Signal Processing Society with multi class CNN classifier.