Secure management of retinal imaging based on deep learning, zero-watermarking and reversible data hiding

Garcia-Nonoal, Zaira; Mata-Mendoza, David; Cedillo-Hernandez, Manuel; Nakano-Miyatake, Mariko

doi:10.1007/s00371-023-02778-1

Cited by 6 publications

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The suggested method in [15] maintains perceptual transparency for a range of Fundus images by embedding patient identification in Fundus image in singular value decomposition domain with adaptive quantization parameter. In study of [16], a safe management system for retinal imaging is proposed. It uses zero-watermarking, reversible data concealing, and deep learning.…”

Section: Introductionmentioning

confidence: 99%

A new highly fractal 1D- chaotic map-based novel encryption on HVS-based watermarked Fundus images for 2-level security in teleophthalmology applications

Amrutha,

Kannammal

2023

Phys. Scr.

View full text Add to dashboard Cite

In recent years the combination of medical image watermarking and encryption is becoming popular due to a better trade-off between authenticity and security, since, each of these methods alone is insufficient to offer complete security. In the proposed work,the watermark is embedded into the Fundus image using the Singular Value Decomposition (SVD) of the La*b* space image of the Fundus image. The scaling factor is selected based on the HVS property of the human eye. The proposed embedding algorithm is semi-blind because the system requires the singular values (s) of the La*b* image at the receiver side in order to extract the watermark. To improve security, the authenticated image then undergoes 2-levels of confusion process followed by diffusion for encryption. For diffusion, the key is generated from the newly developed Modified Sine-Logistic Chaotic System (MSLCS), which is constructed by the combination of logistic map and sine map. The performance evaluation of the new chaotic system has been done in terms of the Bifurcation diagram, Lyapunov exponent, Phase portrait and Entropy. Also, Different tests such as Blood vessel segmentationand Retinal object identification were carried out on the original image, watermarked image, decrypted image and recovered image. The results show that the embedding of watermarks followed by encryption, on Fundus images does not affect the diagnosis and identification of retinal objects. Correct decryption, secured recovery of Fundus image and guaranteed extraction ofthe watermark at the receiver side make the proposed system secure for Fundus images in tele-ophthalmology applications.

show abstract

Section: Introductionmentioning

confidence: 99%