![]() | Only 14 pages are availabe for public view |
Abstract Dealing with digital images today has become available from previous periods with the tremendous development in image editing programs and high-resolution imaging devices with the availability of modern computers. Digital images are of legal and scientific importance in various fields. The field of verifying the integrity of images and detecting traces of tampering with them is an ever-increasing area of research to reduce the attempt to falsify images. The aim of this study is to develop an intelligent electronic system for detecting forgery in digital images. This system helps in the scientific field, criminal, and media research. The study relied on a global image database available for scientific use that includes original and forged images. The main steps of the proposed system include entering the images, whether original or forged, then extracting the features from the digital images through three methods: first, Gray-Level Co-Occurrence Matrix (GLCM); second: Statistical Moments; and finally, combining the features both GLCM and Statistical Moments in order to detect forgery of digital images. Euclidean distance was used to classifying the test image is original or fake image. The proposed system had a high level of performance accuracy when using hybrid GLCM and Statistical Moments. shows a comparison between the three methods used to extract traits through the similarity ratio that was calculated by the Euclidean distance. Therefore, we find that the similarity results are very high resulting from the combination of the two feature vectors for both GLCM and Statistical Moments were better than using the Euclidean distance on the feature vector for both GLCM and Statistical Moments separately. |