Enhanced q-logarithm Based Image Watermarking Using Fuzzy Image Filter

Watermarking is important for copyright protection, preventing unauthorized use of content, and establishing ownership, particularly in digital format such as images and videos. Robust image watermarking aims to embed a watermark that can survive various intentional or unintentional attacks, e.g. noise addition, compression, and geometric transformations, while maintaining the perceptual quality of the original image. In this paper, an improved version of a robust image watermarking method based on the q-logarithm transform is proposed. With a new prediction model of original q-logarithm component based on fuzzy image filtering technique, the embedded watermark can be blindly extracted, and withstand more manipulation and attacks. The proposed watermarking method is tested using two categories of images, i.e. photographic and computer-generated images, and its performance is evaluated using Bit Correction Rate (BCR) and Normalized Correlation (NC) metrics. The results at various wPSNRs and common attacks are presented and compared to the two previous watermarking methods [5] and [6] to demonstrate its superior performance.