Image transformation into device dependent color printer description using 4th-order polynomial regression and LabView object oriented programming development of image processing modules

This thesis investigates the development of printer device profiles used in color document printing system environments when devices with intrinsically different gamut capabilities communicate with one another in a common (CIELAB) color space. While the main thrust of this activity focuses on the output printer, namely the Xerox 5760 printer, and its rendition of some device independent image description, characterizations are provided which investigate relative areas of photographic, monitor, and printer gamuts using a visual hue leaf comparison between devices at standard hue angles determined from a Kodak photographic Q60C. The printer is modeled using 4th-order polynomial regression which maps the device independent CIELAB image representation into device dependent printer CMYK. This technique results in 1.57 AEavg over the 360 training data set. Some key properties of the proposed calibration method are as follows: Linearized CMYK tone reproduction curves with respect to AEpaper to improve the distribution of calibration data in color space. Application of GCR strategy and linearization to the calibration target but perform the regression on the measured CIELAB and original CMY values. Each GCR strategy does not remove CMY so that printer gamut is maximized. This method relies on the regression to determine the appropriate CMY removal. Four GCR strategies were explored: Colorimetric analysis indicates that the 4th order correction plus printer stability accounts for 2.5 to 3.2 AEavg over 200 in gamut colors in the Q60 over the four GCR strategies. Approximately 1 .0 AEavg was attributed to scanner calibration and an additional 1.0 AEavg for out of gamut colors. A library of image processing algorithms is included, using LabView object oriented programming, which provides a modular approach for key color processing tasks. In the user interface, an image is selected with appropriate GCR stategy, and the program operates on the image according to the path displayed below: In general, the pictorial image quality is excellent for each GCR strategy with very subtle differences between GCR approaches. The parabolic GCR offered the best technical attributes in the areas of gamut size and image graininess. Having defined an algorithmic approach to develop and test different image processing approaches, several advanced topics are discussed which investigate out of gamut colors. An example using a gamut alarm is applied to the image set to visually assess out of gamut colors. Included is an outline of a gamut mapping strategy which preserves hue and lightness and performs piece-wise chroma remapping into a printer gamut. The project concludes with an example of how to build a printer characterization so that it conforms to the InterColor output device profile specification