Two problems in digital color imaging: Colorimetry and image fidelity assessor

Wencheng Wu, Purdue University

Abstract

In this research, we investigated two important issues in digital color imaging: the measure of colors and the measure of perceived image fidelity. For measuring colors, we explored the use of a digital camera for imaging colorimetry. Our system consists of a digital camera, a set of filters, and a set of calibration matrices which are illuminant-dependent. We develop two colorimetry systems by applying model-based and regression-based techniques. For the model-based system, a model for the digital camera is employed; and our objective is to find the optimal filters and the corresponding calibration sets that minimize a cost function which accounts for errors in L *a*b* space, system robustness, and filter smoothness. For the regression-based system, no modeling technique is applied to the camera and filters. The objective is simply to find the optimal calibration matrices that minimize the total least squared errors of a given color set in CIE XYZ coordinates under several pre-selected illumination conditions. We apply both types of colorimetry systems to two specific tasks: general purpose measurement of color samples and colorimetry of human teeth. Even though each system has its strength and weakness, we found that both systems yield accurate measure of colors from our experimental results. For evaluating perceived image fidelity, we develop the Color Image Fidelity Assessor (CIFA) which extended Taylor et al's achromatic IFA. The CIFA is a visual model that assesses perceived image fidelity along three opponent coordinates: the luminance, the red-green opponent, and the blue-yellow opponent. In this work, we introduce a novel color descriptor namely chromatic difference to measure the spatial interaction between colors along chromatic coordinates. A set of chromatic difference discrimination experiments that use physiologically motivated stimuli were presented. The experiments were designed in such way that they measured the sensitivity of each visual channels of human visual system that are directly related to the model. Therefore, we can directly apply the results of the psychophysical experiments to CIFA for the task of perceived image fidelity assessment. Our simulation results show that the CIFA provides visual prediction as human perceives for a wide range of image contents and distortion types. Hence we concluded that CIFA is an effective perceived image fidelity assessor.

Degree

Ph.D.

Advisors

Pizlo, Purdue University.

Subject Area

Electrical engineering

COinS