Generation of an ICC profile from a proprietary style file

In color imaging technologies one of the major parts of the scanning procedure is processing the raw RGB scanner data to make it suitable for printing on a given CMYK output device. The processing typically involves applying a number of technical and aesthetic, iteratively developed edits. These edits are stored in a proprietary file format, i.e. a type of vendor-specific macro file. There has been a recent shift from this closed loop type of color control to the open system of color management such as that specified by the International Color Consortium (ICC). ICC color managed workflows can only use a universally agreed file called a profile, and not proprietary user edit files. It may have taken a long time to perfect the proprietary edits and users will not want to lose them when they migrate to the new ICC workflows. This article describes a process that allows users to convert color edit information from a vendor-specific proprietary edit file into an ICC profile. The proposed system works by analyzing the proprietary edit information and then incorporating this information into an input profile. The application of the input profile and an output profile then produces a set of processed CMYK values that, when printed on the target printer produce essentially the same result as that achieved through the proprietary route. As a major part of the process relies on the generation of an accurate input profile, a polynomial fitting process and an empirical process are investigated for use in generating the transform relationship. This research also considers the reversibility of the output profile and a process to ensure accurate reversibility of the output profile is presented. A discussion is presented on the choice of rendering intents available and the different situations where each would be appropriate to use. Experimental results are presented for the whole process based on a test transparency that contains an image of a Macbeth ColorChecker chart. Quantitative results are presented to show how an ICC profiled system can be created that mimics a given proprietary reproduction style to within 2DeltaE. This research demonstrates how novel image engineering applications can be constructed from within the ICC framework.