Forgery detection algorithm based on texture features

Any researcher's goal is to improve detection accuracy with a limited feature vector dimension. Therefore, in this paper, we attempt find and discover the best types of texture features classifiers that are appropriate for coarse mesh finite differenc (CMFD). Segmentation-based fractal analysis (SFTA), local binary pattern (LBP), Haralick have been chosen. K-nearest neighbors (KNN), naïve Bayes, Logistics also among SFTA, fed KNN, logistics classifier. The outcomes experiment indicate SFTA surpassed all other classifiers, making it use forgery detection. has second-best performance classifiers. using LBP lower than features. It shows KNN classifier outperformed two terms accuracy. However, logistic had lowest proposed based method compared state-of-the-art techniques dimension outperforms current techniques.