Robust Image Authentication Using Optimized Haralick Features Based on Genetic Algorithm

Abstract

Recent development in multimedia data increases the role of digital images in many applications. As a result,content based image authentication with an aid of feature extraction techniques has a great impact. It suffers a serious drawback of increased computational complexity due to the availability of irrelevant and interdependent features, which contains no useful information about the image. In the proposed authentication system to decrease the complexity, not all haralick features are used but only the most influencing features that contain critical information about the image is used in hash generation. Feature selection based on chaotic genetic feature selection optimization algorithm is used to optimize the haralick features without compromising the accuracy of the system. A comparison on the system performance with all the features and with optimized features is performed. Experiment results shows resultsusing optimized features are similar to the results of using all the features..