Wavelet based non-informative code for security of fundus images

Abstract

This paper proposes an algorithm for generating unique and non-informative identification code for security of medical fundus images in tele-medicine applications. Unlike the conventional methods like watermarking, the proposed technique does not tamper the medical images which ensure no loss of medical information. The unique digital pattern is generated from strategically combination of statistical features of fundus image in wavelet domain and digital patient ID. The Patient data can be recovered accurately only on access of generated unique digital pattern to solve the security issue of fundus images. The proposed algorithm is experimented on publically accessible DRlVE dataset of medical fundus image and attained promising results in identification of fundus images. The presented work can therefore be implemented in telemedicine industry and its applications where precise and loss-less verification is mandatory as this may have a huge impact at times when human life is at stake.