Digital watermarking on recolored images for deuteranopia
Thitiporn Pramoun, Thumrongrat Amornraksa · Proceedings of the 3rd International Conference on Communication and Information Processing · 2017
Normal color vision people can see all shades in color images while the color blind people cannot. This causes an important problem in the color blind world. To encounter it, we use the image recoloring techniques to help improve the perception of all objects within the images. In digital image watermarking, a watermark embedded within a recolored image will affect its pixel values, and as a result lead to some distortions in color shades appeared to color blind people. In this paper, an image watermarking method for recolored images is proposed so that the watermarked recolored image will still be as effective to color blind people as the recolored one. In the embedding process, the watermark is embedded in Medium (M) wavelengths component of the LMS color space of a recolored image. The watermark strength is tuned by using the luminance component from the recolored image. The watermark is extracted without requiring the original recolored image by exploiting relationship between the predicted M component and the watermarked component within a local area. The watermark robustness against some common attacks and Stirmark is evaluated in terms of weight peak signal-to-noise ratio (wPSNR) and normal correlation (NC), and compared to the previous work.