Research On The Color Mixing And Embedding Of Infrared Detectable Watermark

This thesis design the detectable watermark based on the infraredspectral characteristics of the ink. The watermark is hidden under visible lightand is detected under infrared light. Since no special inks and special printingprocess are needed, it greatly reduces the printing cost and complexity, and itis easy to detect.Firstly, the background and significance of anti-counterfeiting printingtechnology is introduced. Through the analysis of current research the severaldevelopment direction of printing security are described. This thesis hasresearched on the ordinary ink-based infrared detectable watermarks, and it isdescribed in detail. In the theoretical analysis section, the overall designprocess is given, and the core theory in the thesis is discussed. Thesis iscomposed to color space conversion, gray component substitution andhalftone screening. The third chapter discussed the conversion of color spacebased on the artificial neural network, designed depth architecture modelbased on a simple network. The neural network is trained using the collectedcolor card data as the experimental data source and the network performanceis analyzed and evaluated. According to the transform relationship betweenCMYK and LAB color space, color mixing algorithm based on GCR is designed.The watermark is embedded with the halftone screening algorithm. Finally, theactual effects are analyzed and evaluated through printing output anddetection with infrared equipment.