Abstract
Recently various software have been invented in order to make the job of image manipulation more comfortable and effective. Using software images are merged in such a way that ocular review cannot differentiate the resulting forgeries from authentic images. This paper proposes a model for the identification of such merged images. Proposed methodology for identification is the connection of two stagesfeature extraction using principle component analysis and the use of classifier based on the evolutionary artificial neural network (EANN). The behavioral aspects of EANN phenotype was evaluated under changing parameters. The results obtained after experimentation were promising.