KAZE and wild horse optimization based copy-move forgery detection of fundas images

The goal of modern health information systems is to digitise medical records, which include fund as pictures, to deliver high-quality patient care. It is necessary to perform forgery detection on fund as images prior to treating patients because the images can be manipulated through copy-move forgery with the malicious goal of hiding some lesions or creating multiple copies of lesions. This can result in wrong treatments and expensive surgical procedures that could cause infections and blindness in the patients. The current forgery detection techniques might not yield adequate results since the key points are not evenly dispersed throughout the full image area. The goal of this research is to provide an effective approach for forgery detection that distributes the key-points over the entire image regions, uses speeded-up robust features to assess the features at key-points, performs feature clustering using wild-horse optimisation, and carries out feature matching and false match removal. This paper demonstrates the superiority of the proposed method by analysing and comparing the performances such as accuracy, sensitivity, specificity, and precision, of the suggested technique with those of existing methods on 1440 fund as images.