Peter Nussbaum

Using gaze information to improve image difference metrics

We have used image difference metrics to measure the quality of a set of images to know how well they predict perceived image difference. We carried out a psychophysical experiment with 25 observers along with a recording of the observers gaze position. The image difference metrics used were CIELAB ΔEab, S-CIELAB, the hue angle algorithm, iCAM and SSIM. A frequency map from the eye tracker data was applied as a weighting to the image difference metrics. The results indicate an improvement in correlation between the predicted image difference and the perceived image difference.

Regression based characterization of color measurement instruments in printing applications

In the context of print quality and process control colorimetric parameters and tolerance values are clearly defined. Calibration procedures are well defined for color measurement instruments in printing workflows. Still, using more than one color measurement instrument measuring the same color wedge can produce clearly different results due to random and systematic errors of the instruments. In certain situations where one instrument gives values which are just inside the given tolerances and another measurement instrument produces values which exceed the predefined tolerance parameters, the question arises whether the print or proof is approved or not accepted with regards to the standard parameters. The aim of this paper was to determine an appropriate model to characterize color measurement instruments for printing applications in order to improve the colorimetric performance and hence the inter-instrument agreement. The method proposed is derived from color image acquisition device characterization methods which have been applied by performing polynomial regression with a least square technique. Six commercial color measurement instruments were used for measuring color patches of a control color wedge on three different types of paper substrates. The characterization functions were derived using least square polynomial regression, based on the training set of 14 BCRA tiles colorimetric reference values and the corresponding colorimetric measurements obtained by the measurement instruments. The derived functions were then used to correct the colorimetric values of test sets of 46 measurements of the color control wedge patches. The corrected measurement results obtained from the applied regression model was then used as the starting point with which the corrected measurements from other instruments were compared to find the most appropriate polynomial, which results in the least color difference. The obtained results demonstrate that the proposed regression method works remarkably well with a range of different color measurement instruments used on three types of substrates. Finally, by extending the training set from 14 samples to 38 samples the obtained results clearly indicate that the model is robust.

Color Image Quality in Projection Displays: a case study

Recently the use of projection displays has increased dramatically in different applications such as digital cinema, home theatre, and business and educational presentations. Even if the color image quality of these devices has improved significantly over the years, it is still a common situation for users of projection displays that the projected colors differ significantly from the intended ones. This study presented in this paper attempts to analyze the color image quality of a large set of projection display devices, particularly investigating the variations in color reproduction. As a case study, a set of 14 projectors (LCD and DLP technology) at Gjovik University College have been tested under four different conditions: dark and light room, with and without using an ICC-profile. To find out more about the importance of the illumination conditions in a room, and the degree of improvement when using an ICC-profile, the results from the measurements was processed and analyzed. Eye-One Beamer from GretagMacbeth was used to make the profiles. The color image quality was evaluated both visually and by color difference calculations. The results from the analysis indicated large visual and colorimetric differences between the projectors. Our DLP projectors have generally smaller color gamut than LCD projectors. The color gamuts of older projectors are significantly smaller than that of newer ones. The amount of ambient light reaching the screen is of great importance for the visual impression. If too much reflections and other ambient light reaches the screen, the projected image gets pale and has low contrast. When using a profile, the differences in colors between the projectors gets smaller and the colors appears more correct. For one device, the average ΔE*ab color difference when compared to a relative white reference was reduced from 22 to 11, for another from 13 to 6. Blue colors have the largest variations among the projection displays and makes them therefore harder to predict.

Effect of time spacing on the perceived color

One of latest developments for pre-press applications is the concept of soft proofing, which aims to provide an accurate preview on a monitor of how the final document will appear once it is printed. At the core of this concept is the problem of identifying, for any printed color, the most similar color the monitor can display. This problem is made difficult by such factors as varying viewing conditions, color gamut limitations, or the less studied time spacing. Color matching experiments are usually done by examining samples viewed simultaneously. However, in soft proofing applications, the proof and the print are not always viewed together. This paper attempts to shed more light on the difference between simultaneous and time-spaced color matching, in order to contribute to improving the accuracy of soft proofs. A color matching experiment setup has been established in which observers were asked to match a color patch displayed on a LCD monitor, by adjusting its RGB values, to another color patch printed out on paper. In the first part of the experiment the two colors were viewed simultaneously. In the second part, the observers were asked to produce the match according to a previously memorized color. According to the obtained results, the color appearance attributes lightness and chroma were the most difficult components for the observers to remember, generating huge differences with the simultaneous match, whereas hue was the component which varied the least. This indicates that for soft proofing, getting the hues right is of primordial importance.

Evaluating an image based multi-angle measurement setup using different reflection models

Image based measurement setups are widely used for multidirectional reflectance measurements of materials. Different reflection models are used to estimate the material reflectance. In this paper we use two commonly known simple reflection models to evaluate an image based measurement setup proposed in our previous studies. Sample material is measured at multiple incident and viewing directions using the setup. The captured data is divided into training and test set to evaluate the setup. Two reflection models are trained using the training dataset. To evaluate the measurement setup, one of the sample materials is measured at few incident and viewing directions using a light source and a Tele-Spectro-Radiometer (TSR). This measured data is then compared with the reflectance estimated by the two reflection models. Results shows that image based multi-directional reflectance measurements can be performed using measurement setup proposed in our previous work. The data captured using the setup can be used to fit different reflections models for the sample materials used in this setup. Introduction Image based instruments that are relatively cheap and fast, are being increasingly used in multi-angle reflectance measurements of the material [1, 2, 3]. The multi-angle data captured can be further processed (for example to estimate the material BRDF). Where the need is to acquire fast and cheap measurements, image based instruments can be a good alternative compared to the expensive, time consuming, but precise instruments (like Goniospectrometers). These measurements can then be used to train reflection models to calculate the BRDF and simulate or reproduce the material appearance. A number of reflection models are proposed till now which mainly aim towards fulfilling the needs of the computer graphics field to simulate material appearance in a given scence/situation. One of the main aims of the computer graphics field is to accurately measure the appearance of materials and simulate/reproduce synthetically the real object using simple methods/techniques [4]. A Bidirectional Reflection Distribution Function (BRDF) is a distribution function that describes the surface reflectance properties of opaque and homogenous materials [5] as given in Equation (1). fBRDF (θi,φi;θr,φr,λ ) = dLr(θi,φi,θr,φr,Ei) dEi(θi,φr) (1) Here, i and r denote incidence and reflection respectively. θ and φ together indicate the direction, Ei is incident irradiance, Lr is radiance and d is the differential. Guarnera [4] presented an overview of the BRDF models used to represent surface/material reflection characteristics. As discussed in [4], BRDF models can be classified into Physically-based models and Phenomenological models. Physically-based models are based on physics and optics and are described using micro-facets of varying size and orientations. One of the very well known and widely used physical models is the Cook-Torrance [6] model. Phenomenological models are approximations of the reflectance data using measured data and fitting of the same using analytical models. Some of the commonly known phenomenological models are the Phong [7], Ward [8], Lafortune [9], etc. In our previous studies [3] we presented a measurement setup which can perform multi-angle reflectance measurements of homogenous, flexible materials in a fast and relatively cheap way. In this setup, we mount the flexible sample onto a cylinder of known radius and measure using a RGB camera [10, 3]. The captured RGB data is converted into the colorimetric space CIEXYZ using the conversion matrix M̂. Matrix M̂ is derived using the camera spectral sensitivity (measured using a monochromator) and the CIE 2◦ colour matching functions by minimising the error using least square technique [11]. This data can be used further as training data to train different reflection models (like Cook-torrance or Ward model) to estimate the sample BRDF and simulate the material appearance. The objectives of the work presented in this paper are: • to estimate the surface reflectance properties of the sample materials used in the experimental setup using the image based multi-angle measurement setup proposed in previous studies [3], • to train different reflection models using the colorimetric data captured with the image based multi-angle measurement setup and evaluate the setup against measurements obtained using a TSR, • test the trained reflection models to predict the sample material reflectance at different illumination and viewing directions. Background In order to evaluate the measurement setup we used 2 reflections models; Cook-Torrance (hereby referred as CT in this paper) and Ward to fit the measurement data obtained in the measurement setup by using the RGB camera [3]. The trained model parameters were then used to estimate the BRDF measurement at different incident and viewing directions. CT model is a physical model that describes the intensity and spectral composition of the light reflected from the object/material. CT model as described in

Quality of color coding in maps for color deficient observers

For a color deficient observer, the quality of a map or other information design may be defined as the ability to extract features. As color is such important conveyor of information, the colors need to appear correct and be perceived in the desired and intended way. As color appearance is affected by the size of the stimuli, the task of discriminate colors may be even more difficult for a color vision deficient observer. In order to investigate the discriminability of the color coding in an official Norwegian map product, we conducted an experiment involving both color deficient and color normal observers. Also, we investigate to what extent the ability to discriminate colors is influenced by size of the visual field. The experiment revealed that the color vision deficient observers made significant more errors than the normal observers, especially when the visual angle was reduced. Introduction It is established that color is undoubted guide to visual attention [1]. Color vision deficiency (CVD) is a condition that affects the ability to distinguish and discriminate colors, and it is believed that about 8% of all males and 0,5 % females is affected by such condition [2, 3]. Studies reviewed in the work of Cole [4] reveal that up to 60% of the users with abnormal color vision reported problems in observational task as reading color coded charts, slides and prints. Further, they are slower and make more errors in search when color is an attribute of the target object or is used to organize the visual display. It is well known that the color appearance is affected by size of the stimuli, which is often referred to as the color size effect. Carter and Silverstein [5] state that a common experience is difficulty discriminating the multiple small-subtense colors on weather, financial, and other maps and charts viewed on print media and conventional-size information displays. Color is a cartographic element that is used both aesthetically and as a conveyor of information and is well documented in literature [6-9]. Color is the cartographic element that is most frequently misused and is often considered as a difficult cartographic element to use, as it easily draws attention away from the data and goals for the map when it is used poorly [8]. Based on the assumption that quality of a map can be based on the ability for a user to extract features, the main goal in this study is to investigate the following research questions:  Are the colors in an official Norwegian map product distinguishable for a color deficient observer?  Is the ability to distinguish or recognize colors more influenced by variation of visual angle for color deficient observers compared to observers with normal color vision? The paper is organized as follows: First, an overview of basic concepts and related work necessary to understand the research topic is given. Then the experimental setup is described, followed by results, discussion and conclusion. Background and related work This section gives an overview of some of the basic concepts and related work which are essential for the understanding of this research topic. An overview on cartography and color coding, color vision deficiencies, color size effect and visual field is provided in the following pages. Maps and cartography In reference maps, hue is used to symbolize different kinds of features like the blue and green for water and land. Color may be used to represent qualities or quantities and often in a combination, like a topographic map will use hue in order to present qualitative data (like water) or classify data (like class of roads) and value to classify and represent quantitative data (like deepness of the water). Bertin [9] states that “Above all, color exercises an undeniable psychological attraction. In contrast to black-and-white it is richer in cerebral stimulation, and in numerous cases where it can appear as a luxury, this luxury nonetheless “pays off”. It captures and holds attention, multiplies the number of readers, assures better retention of the information, and, in short, increases the scope of the message. Color is particularly applicable to graphic messages of a pedagogic nature”. The use of color in maps is most based on conventions and traditions. Some of the conventions are to choose colors that have a similarity to real life objects, like green and blue to represent land and water. Other conventions are to use strong colors to emphasize important objects, like the use of red to represent highways or cities. Standards or specifications exists for thematic maps in different industries like geology [10, 11], hydrographic services and nautical charts [12] and sports like orienteering [13, 14]. There is also a standard from the International Standardization Organization, ISO 19117:2012 Geographic informations Portrayal [15]. This standard handles presentation of geographic data, but the focus is more on symbolization than use of colors. Then there are the national standards or specifications (often adopting on the ISO 19117), like the standard of the Norwegian Mapping Authority [16] and the specification for cartography on digital displays [17]. Color vision and color deficiency Trichromacy is the ability to perceive colors through the three types of cones (L, M and S) that respond to different wavelengths of the visible spectrum. People with full color vision perceive color with all three cones, while people with different color vision anomalies are affected by either a lack of or dysfunction of one or more of the cones. Color vision deficiencies are most often a congenital condition, but may also be caused by medical conditions. Color vision deficiencies are well described in literature, for example in the book by Hansen [2]. As for the classification of the color vision it is categorized according to available cones (monochromatism, dichromatism and anomalous trichromatism) and further categorized by the cones that are missing or dysfunctional:  Protanopia or Protanomaly (missing or dysfunctional L-cone (red)).  Deuteranopia or Deuteranomaly (missing or dysfunctional Mcone (green)).  Tritanopia or Tritanomaly (missing or dysfunctional S-cone (blue)). A missing or dysfunctional cone does not make the observer “color blind”, but the color perception is limited and the observer may have problems perceiving certain colors accurately and may confuse color ranges. Image enhancement methods for color vision deficient observers A typical method for enhancing images to improve their quality for CVD observers [18] is to change the color (re-coloring methods). Specializations of such methods are the daltonization methods, aiming to compensate poor color discrimination by targeting a specific type of CVD. An overview of existing methods and evaluation of these by measuring accuracy and response time in sample-to-match and visual search methods involving CVD are well described in the work of Simon-Liedtke et al. [19]. Other examples of re-coloring methods is the palette method proposed by Green [20] and a version of the palette method applied to maps proposed by Kvitle et al. [21]. Simulations of different color deficiencies and daltonization methods have been applied in order to create better maps for the CVD observer. The ColorBrewer [7, 22] and Colororacle [23], both based on algorithms like those described in Brettel et al. [24] are examples of applications dedicated to maps. One major problem with simulations is that they operate more like a guideline for a designer with normal color vision. The ColorBrewer is evaluated by CVD observers in experiments involving multiple choice map reading questions as described in the work of Gardner [25], but in a limited group of subjects. The color palettes described in ColorBrewer are well suited for thematic mapping such as choropleth maps, but are not directly applicable to the common reference map (which is the aim of our study) because of the limited set of colors. The Ordnance Survey [26] has developed new reference maps for CVD observers, where the maps are actually developed in an iterative way in cooperation with CVD observers. The group of ten CVD observers was recruited among employees, and these users contributed in several phases of the development. These maps have been embraced by both CVD and color normal observers, even though some of the color coding are unconventional. Color size effect and visual angle The visual angle [27] or the objects angular size is the relationship between the viewing distance and the size of the object. Years of research states that color appearance changes according to different sizes of the same color stimulus. An overview of history of this research and a structural model is given in the work of Carter and Silverstein [5]. An attempt to take this effect into account in the CIECAM02 color appearance model is described by Luo and Li [28] and experimental work on variation of size on display are described in Xiao et al. [29]. Exploring the color size effect for CVD observers, like the work of Pokorny and Smith [30], Nagy et al. [31] and Paramai et al. [32] indicate a correlation between stimulus size and performance in color-naming tasks as many dichromats show trichromacy in color matching and name the colors in fair agreement with the normal observers in larger fields of view (≥3°). Also, the experiments [31] revealed large individual difference between the observers and that for some observers the ability to discriminate varied considerably as a function of the stimulus parameters, such as field size, luminance level and method of presentation. As summarized in Pokorny and Smith [30], in large fields dichromats show weak residual trichromatic vision and at higher luminances deuteranopes and protanopes show residual deuteranomaly and protanomaly. Experimental work Based on the goals to test if the colors of a known official map are discriminable for CVD observers, and if color size

Adaptive color rendering of maps for users with color vision deficiencies

A map is an information design object for which canonical colors for the most common elements are well established. For a CVD observer, it may be difficult to discriminate between such elements - for example, it may be hard to distinguish a red road from a green landscape on the basis of color alone. We address this problem through an adaptive color schema in which the conspicuity of elements in a map to the individual user is maximized. This paper outlines a method to perform adaptive color rendering of map information for users with color vision deficiencies. The palette selection method is based on a pseudo-color palette generation technique which constrains colors to those which lie on the boundary of a reference object color gamut. A user performs a color vision discrimination task, and based on the results of the test, a palette of colors is selected using the pseudo-color palette generation method. This ensures that the perceived difference between palette elements is high but which retains the canonical color of well-known elements as far as possible. We show examples of color palettes computed for a selection of normal and CVD observers, together with maps rendered using these palettes.

Verification of proposed ISO methods to measure resolution capabilities of printing systems

In this paper we deal with a new Technical Specification providing a method for the objective measurement of print quality characteristics that contribute to perceived printer resolution, “ISO/IEC TS 29112:2012: Information Technology – Office equipment – Test charts and Methods for Measuring Monochrome Printer Resolution”. The Technical Specification aims at electrophotography monochrome printing systems. Since the referred measures should show system or technology independence inkjet printing systems are included as well in our study. In order to verify if given objective methods correlate well with human perception, a psychophysical experiment has been conducted, and the objective methods have been compared against the perceptual data.