An enhanced copy-move forgery detection using machine learning based hybrid optimization model

In recent years, one of the forgery detection methods is the Copy-Move Forgery Detection (CMFD), which is mostly used approach amongst all the forgery identification techniques. In some forgery identification schemes, it is rigid to distinguish falsified regions or real similar regions. However, detection of the tampered region is very difficult if the interacted area is rescaled or rotated in the existing tampering recognition algorithms. Hence, an Enhanced CMFD approach (ECMFD) is projected in this paper to tackle the difficulties experienced by the existing approaches. The proposed technique developed an innovative and efficient Machine Learning (ML) approach called Prewitt mm(') Filter with Generalized Action Selection based Hybrid Artificial Bee Colony African Buffalo Optimization (PFGAS based HABC-ABO). Initially, the proposed PF model is utilized for de-blurring and detecting the edges of the input images. Subsequently, the de-blurred images are separated into blocks using GAS, and the copy-move forged region in digital images under different scenarios such as scaling and rotation is detected using HABC-ABO. In this approach, a portion of the picture is copied and localized in another region in the corresponding picture for including or screening some valuable information that exists in the image. The simulation of this method is done with the use of Python, and the performance analysis indicates that the proposed PFGAS based HABC-ABO method has a better effect on rotation and scaling. Hence, it attained high performance in parameter metrics like accuracy, precision, and recall than other existing approaches.