Feature Extraction-based Medical Image Watermarking Using Fuzzy-based Median Filter

Abstract

In recent days, securing the clinical images with high-level attacks is still a great challenge. The proposed algorithm focuses on developing an algorithm, which starts with pre-processing clinical image by Switched Mode Fuzzy Median Filter (SMFMF) for extracting the noise affected pixel thereby replacing the noisy pixel with the value of median pixel. Clinical image features, such as colour, shape, and texture, are extracted for selecting appropriate region to watermark, which increases the process of watermarking and decreases the algorithm’s computational time. At watermark, calculating the key component can increase the security level and source clinical image singular values and singular values of watermark images are modified to review the robustness of clinical image. At the watermark extraction point, watermark validation for permanent checks is achieved via the logistic map technique to unravel the watermark. The peak noise signal ratio (PSNR) and normalized cross correlation values (NCC) measure the clinical image’s multi-target wellness predictions. In addition to watermarking validation, the proposed algorithm offers improved clinical strength and protection.