基于热力图和 SVD 的数字媒体抗摄屏水印技术研究OA

To overcome the limitations of current watermarking technologies,such as fixed embedding strength and a low peak signal-to-noise ratio(PSNR),this paper presents a screen-capture-resistant watermarking method for digital media.This method is based on singular value decomposition(SVD)and heat maps.User interaction data with digital media is collected using FineReport,and user behavior features are extracted and user attention levels are calculated to generate digital media heat maps.These maps are then used to determine suitable watermark embedding areas and compute adaptive watermark embedding strength.An improved SVD algorithm decomposes the selected embedding region.The jointly encoded watermark information and its corresponding embedding strength are incrementally embedded through additive operations.After reconstruction,digital media embedded with the screen-capture-resistant watermark are obtained.Experimental results demonstrate that the embedding region identified by the proposed method aligns with the optimal embedding area and that the embedding strength corresponds to actual requirements.The maximum average PSNR achieved is 98.