Digital tags from multi-subject biometric features for joint ownership of digital data

Abstract

This paper addresses the issue of joint ownership of digital images and audio signals by embedding a perceptually transparent digital signature. The pattern is generated using an approach so that the authentication of each constituent subject can be done after decomposition of the digital pattern. The features from these iris images from multiple subjects are strategically combined to generate a biometric watermark of an acceptable size under perceptual transparency constraints. This biometric watermark is embedded in SVD domain for audio signal and using a DCT-based method for digital images. The extracted watermark is decomposed into the respective iris templates for distinct identification in the database. Experimental result shows that proposed watermarking system maintains the challenging design requirements of watermarking system. Biometric watermark generated from biometric features of more than one subject is a novel effort for establishing joint ownership of multimedia data and may be considered as a significant contribution.