Forged Image Detection using SOTA Image Classification Deep Learning Methods for Image Forensics with Error Level Analysis

The area of computer vision [1] has excelled in terms of innovation and performance delivered by leveraging Deep Learning [2]. The various tasks of computer vision are classification, object detection [3], object counting [4], image segmentation [5, 6] and many more. Classification [7] can be termed as one of the most primordial tasks in computer vision. The task of classification is identification of an entity or object by prognosticating its appropriate label is done effectively using Convolutional Neural Networks [8] abbreviated as CNN. The standard CNN gone under massive improvements in the latter period when ImageNet [9] Large Scale Visual Recognition Challenge (ILSVRC) [10] came into existence yielding many models that are currently state-of-the-art deep learning models for image classification.