William B. Cowan

Color gamut mapping and the printing of digital color images

Principles and techniques useful for calibrated color reproduction are defined. These results are derived from a project to take digital images designed on a variety of different color monitors and accurately reproduce them in a journal using digital offset printing. Most of the images printed were reproduced without access to the image as viewed in its original form; the color specification was derived entirely from calorimetric specification. The techniques described here are not specific to offset printing and can be applied equally well to other digital color devices. The reproduction system described is calibrated using CIE tristimulus values. An image is represented as a set of three-dimensional points, and the color output device as a three-dimensional solid surrounding the set of all reproducible colors for that device, called its gamut. The shapes of the monitor and the printer gamuts are very different, so it is necessary to transform the image points to fit into the destination gamut, a process we call gamut mapping. This paper describes the principles that control gamut mapping. Included also are some details on monitor and printer calibration, and a brief description of how digital halftone screens for offset printing are prepared.

An experimental comparison of RGB, YIQ, LAB, HSV, and opponent color models

The increasing availability of affordable color raster graphics displays has made it important to develop a better understanding of how color can be used effectively in an interactive environment. Most contemporary graphics displays offer a choice of some 16 million colors; the users problem is to find the right color. Folklore has it that the RGB color space arising naturally from color display hardware is user-hostile and that other color models such as the HSV scheme are preferable. Until now there has been virtually no experimental evidence addressing this point. We describe a color matching experiment in which subjects used one of two tablet-based input techniques, interfaced through one of five color models, to interactively match target colors displayed on a CRT. The data collected show small but significant differences between models in the ability of subjects to match the five target colors used in this experiment. Subjects using the RGB color model matched quickly but inaccurately compared with those using the other models. The largest speed difference occurred during the early convergence phase of matching. Users of the HSV color model were the slowest in this experiment, both during the convergence phase and in total time to match, but were relatively accurate. There was less variation in performance during the second refinement phase of a match than during the convergence phase. Two-dimensional use of the tablet resulted in faster but less accurate performance than did strictly one-dimensional usage. Significant learning occurred for users of the Opponent, YIQ, LAB, and HSV color models, and not for users of the RGB color model.

Changes in perceived color due to chromatic interactions

Studies of chromatic induction have generally examined either (a) the effects of a chromatic surround on a neutral test field, or (b) the effects of one spectral hue on another. To investigate how colors interact in other regions of color space an experiment was designed using fifteen test stimuli scattered through C.I.E. color space. The perceived hue of each stimulus was matched on its own and in the presence of five inducing stimuli. Matching was done both with and without a lens to correct axial chromatic aberration, which was found to be a significant prereceptoral factor influencing perceived colour. With chromatic aberration corrected the overall pattern of chromatic changes can be explained neither by receptor processes alone, nor by opponent channel processes alone. But a reasonable fit can be obtained if changes are allowed to take place in both levels of the system.

Deuteranomalous color matching in the deuteranopic eye

Two observers were classified as deuteranopes by standard tests including two-degree anomaloscope matches. Color matching similar to the Rayleigh type was then carried out for a 10-degree field size at retinal illuminance ranging from 1 to more than 3000 trolands (td). The results show that at the larger field size and higher levels of retinal illuminance, a third independent color-mediating mechanism with the sensitivity of the deuteranomalous cone is participating in the color match. The results also confirm participation of a different third mechanism with rod sensitivity at levels below about 100 td. There is a range of transition between the two as the level increases above 100 td. Therefore large-field color matching in these deuteranopes is trichromatic at the levels tested, not dichromatic, and a third cone system is found to operate at typical photopic light levels under static viewing conditions in a dichromatic eye.

The linear separability effect in color visual search: Ruling out the additive color hypothesis

Bauer, Jolicoeur, and Cowan (1996b, 1996c) demonstrated difficult visual search for color targets that were not linearly separable (in color space) from two distractor colors and easier search for linearly separable targets. This suggested that search is mediated by a chromatically linear discrimination mechanism (see D’Zmura, 1991). However, in those experiments, the targets that were not linearly separable fell midway between the distractor colors and thus corresponded to the admix of the distractor colors. An alternate interpretation of the results of Bauer et al. is that search was more difficult when the target corresponded to the distractor admix than when it did not. We tested this hypothesis in three experiments by contrasting conditions in which a target that was not linearly separable did or did not correspond to the admix of the distractor colors. In all cases, a target that was not linearly separable produced difficult search, demonstrating that linear separability determines search performance.

A practical approach to calculating luminance contrast on a CRT

Luminance contrast is the basis of text legibility, and maintaining luminance contrast is essential for any color selection algorithm. In principle, it can be calculated precisely on a sufficiently well-calibrated display surface, but calibration is very expensive. Consequently, most current systems deal with contrast using heuristics. However, the usual CRT setup puts the display surface into a state that is relatively predictable. Luminance values can be estimated based on this state, and these luminance values have been used to calculate contrast using the Michelson definition. This paper proposes a method for determining the contrast of colored areas displayed on a CRT. It uses a contrast metric that is in wide use in visual psychophysics and shows that the metric can be approximated reasonably without display measurement, as long as it is possible to assume that the CRT has been adjusted according to usual CRT setup standards.

Human Perception of Laid-Out Graphs

Combinatorial graphs are increasingly used for information presentation. They provide high information density and intuitive display of multiple relationships, while offering low cost because they can be created algorithmically. Essential to algorithmic graph layout is a set of rules that encode layout objectives. How these rules are related to inferences drawn from the graph by human observers is a largely unexplored issue. Thus, success or failure by algorithmic standards is only uncertainly related to perceptual effectiveness of the resulting layout. Human experimentation is the only way to correct this deficiency.This poster describes empirical research conducted in 1994. Forty-six respondents, separated into naive and computer-aware groups, freely viewed a collection of graph layouts, providing semantic conclusions they reached on the basis of the layout, in the absence of any semantic attribution to nodes in themselves. We were interested in two questions. First, are semantic attributions consistent or random? If the former semantic objectives must be considered when creating layout rules or objective functions for automated graph layout. Second, if consistent semantic attributions exist, what are they? The remaining paragraphs of this abstract describe our results and conclusions.

Tracking visual search over space and time.

Visual perception consists of early preattentive processing and subsequent attention-demanding processing. Most researchers implicitly treat preattentive processing as a domain-dependent, indivisible stage. We show, however, by interrupting preattentive visual processing of color before its completion, that it can be dissected both temporally and spatially. The experiment depends on changing easy (preattentive) selection into difficult (attention-demanding) selection. We show that although the mechanism subserving preattentive selection completes processing as early as 200 msec after stimulus onset, partial selection information is available well before completion. Furthermore, partial selection occurs first at locations near fixation, spreading radially outward as processing proceeds.

The chromatic cornsweet effect

The Cornsweet effect was measured using equiluminous chromatic gradients as well as with an achromatic gradient. The chromatic Cornsweet effect is smaller than the achromatic effect.

On Clustering in the Location of R&D: Statics and Dynamics

Empirical analyses of research and development find strong evidence that these activities tend to cluster geographically. Clusters are thought to emerge from the presence of localized positive externalities. This paper presents a model of this clustering behaviour. We find that phase changes in clustering exist both as the strength of local externalities changes and as the degree of heterogeneity among firms changes. The dynamics of the system are examined as it responds to shocks to the size of the market for RD followed by a slow re-agglomeration process as producers change their spatial decisions to lower the costs of the new production level.