Journal Article

Pullikolam assisted medical image watermarking on reconfigurable hardware

Sridevi, A, Raj, Vinoth, Bhogi, Tarunraj, Padmaa, M, Siva, Janakiraman, Rengarajan, Amirtharajan, Sivaraman, R, Manikandan, V

2023

Overview

This paper proposes that Pullikolam influenced medical image watermarking on Field Programmable Gate Array (FPGA). This watermarking scheme undergoes three successive operations: Look Up Table (LUT) formation for Pullikolam, Pixel localisation and Least Significant Bit (LSB) substitution. Additionally, this data hiding strategy employs on-chip embedded memory to store the data. This proposed architecture has been designed using Verilog HDL and implemented on Altera Cyclone IV E EP4CE115F29C7 FPGA. This watermarking scheme is consumed 3269 Logic Elements (LEs) which is 3% of the available LEs. Also, it requires only 5.24 mS to perform watermarking on 256 × 256 medical images. This work has achieved an average Mean Square Error (MSE) of < 5, and Normalized Cross-Correlation (NCC) is close to an ideal which is evidence of the robustness of the algorithm. Further, near-zero perceptual transparency has been attained through the Structural Similarity Index Matrix (SSIM) and histogram analyses that confirm the strength of the scheme.

• Publisher: Springer Nature B.V
• Subject: Algorithms / Chips (memory devices) / Cross correlation / Field programmable gate arrays / Lookup tables / Medical imaging / Reconfigurable hardware / Watermarking