Camera Model Identification with Convolutional Neural Network
Komal Mehta Bhagat, Prachi Kaushik, Harsh Nigam · 2023
With the recent developments in multimedia, the forging of images has become an increasing threat for forensics. This study proposes a method using deep learning to trace an image back to its source camera model. In this study the classification performed for a set of 7 camera models but it is extendable and can be used between any numbers of models. It depends on training data given to the algorithm. Two techniques were used for simulations; Dense Neural Networks (DNN) and Convolutional Neural Networks (CNN). Both of these are state of the art techniques yield a strong model. In results the comparison of both the techniques are shown. Result has shown that CNN provides more accurate results.