Development and Calibrating Features for Digitally Forged Image Authentication

D. Subitha Priyadharshini, P. Maya

Abstract: Nowadays image forgeries are done in many ways using software tools. To avoid the image forgery a hash method is been developed which has three secret keys. It is also used to detect image forgery including insertion, removal, object replacement, abnormal colour modification and locating forged areas. Hash values are developed by concatenating local features and global features. For more security skin based biometric values are added with the hash construction. A test image is compared with the hash of a reference image to find whether the image is similar or forged. A network called triple key chaotic neural network is implemented and its a versatile network to make the image secure. And the image will be retrieved and restored.

Keywords: robust hashing, forgery detection, skin based biometrics, chaotic neural network