STUDY AND IMPLEMENTATION OF COPY-MOVE FORGERY DETECTION METHODS IN IMAGE PROCESSING

Abstract

This research work is based on copy-move forgery detection. The copy-move forgery detection technique has various phases which include pre-processing, feature extraction and marking of forgery part on the image. In the previous methodology PCA algorithm was used for the feature reduction in the copy-move forgery detection. The parameters which are taken as input by the PCA algorithm for the feature reduction were defined statically. In this research work, to improve performance of the copy-move forgery detection model, PCA parameters needs to define dynamically. The GLCM algorithm is used for the feature extraction in this research work. The GLCM algorithm extract 13 textural features and mean of all features will be given as input to the PCA algorithm. The initial parameters of PCA algorithm will be defined dynamically for the copymove forgery detection. At last Euclidean distance is calculated between the block pixel to define forgery portion in the image. The performance of the proposed algorithm is tested in terms of precision, recall and F-measure. It is found that proposed algorithm performs better as compared to existing method in terms of all three parameters.