A New Blind Watermarking Technique Based on Independent Component Analysis

Abstract

In this paper, we propose a new blind watermarking scheme by embedding a digital image signature/watermark in the wavelet domain and using Independent Component Analysis (ICA) technique for blind watermark extraction. The watermark is inserted in the middle frequency sub-bands. A visual masking function is used to modify the digital signature to assure the perceptual quality of the watermarked image. Independent Component Analysis (ICA) method is employed for blind watermark extraction without requiring the original image. The presented technique has been successfully evaluated and compared with other wavelet-based blind watermarking techniques against the most prominent attacks, such as JPEG and JPEG2000 compression and quantization.KeywordsIndependent Component AnalysisIndependent Component AnalysisWatermark SchemeJPEG2000 CompressionWatermark ExtractionThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.