Manuscript Details


Title

COLOR IMAGE AUTHENTICATION USING MULTIPLE WATERMARK HIDING AND VARIABLE ENCODING OF SECRET BITS

Authored By

Mr. Soumit Chowdhury

ASSISTANT PROFESSOR OF COMPUTER SCIENCE & ENGINEERING, GOVERNMENT COLLEGE OF ENGINEERING & CERAMIC TECHNOLOGY, KOLKATA


Mr. Divyalok Sinha

Ex-student of Computer Science & Engineering, Govt. College of Engineering & ceramic Technology, Kolkata


Mr. Arkadip Ray

Final Year Student of Computer Science & Engineering, Govt. College of Engineering & Ceramic Technology, Kolkata


Mr. Shanawaz Sams

Part Time Teacher of Computer Science & Engineering, Govt. College of Engineering & Ceramic Technology, KolkatC


Dr. Nabin Ghoshal

Assistant Professor, Department of Engineering & Technological Studies, University of Kalyani

Category

COMPUTER SCIENCE & ENGINEERING (CS)

Abstract

The work highlights a stronger form of color image authentication scheme where four secret color images are invisibly hidden within the different non-overlapping regions of the color cover image for complying robustness and multi-phase authentication. The approach promotesreasonable recovery of at least one of the secretly hidden images under the common image processing operationsby virtue ofthe dynamically variable ranges of bit hiding concepts. Bits of the secret images are fabricated by adopting distinctly different encoding mechanisms for the separate pixel byte elements of a 2X2 sub-image block of the cover image while random circular rotations of these bits within thesub-block elementsbasically ensuring multi type encoding of the secret bits. Mixture of threshold based and redundant typed delicate value adjustment driven bit insertion strategy effectively handles the situation of minor pixel byte value alterations due to the noiseattacks whereas stress is categorically given for more randomized coded values considering the issues of brute force attack. Ultimately the idea injects variable distortionsfor the secret bit fabricated pixel bytes as the effect of the image processing attacks and eventually encouragesacceptable extractions of thesesecretly hidden images. Experimental results have thoroughly confirmed excellent data hiding imperceptibility in every aspect while very little deviation is observed between the original and the fabricated data patterns. Thework clearly establishes superior data hiding imperceptibility as compared to the existing works and even demonstrates good robustness against the image processing operations making the approach suitable for the better authentication scenarios.