Visiblity Optimization for Categorical Data - Term Paper Example

Visibility Optimization for Categorical Data Students Name Name of the Institution 22nd November, 2011. 1.0 CHAPTER ONE: 1.1 INTRODUCTION This project is an application of colors in the real world. This helps to make presentations of a wide range of categorical data types in a more clear way. Categorical variables are represented by a distinct color each. This helps to distinguish between each variable in a visual presentation. However, designers face the problem of selecting optimum colors for each categorical variable for maximum visibility of each one of them. Visual presentations will make the perfect meaning when all categorical data is clearly visible. This project generates an algorithm of selecting colors to represent various distinctive categorical data points. Different variables should have different colors that are agreeable, distinct, presentable and related. For a specific representation say a map, the colors representing vegetation, water bodies, boundaries, population and so on should be visible, separable, salient and distinct. They should also show a relationship that is close to the real life situation. The main emphasis will be to make visible the small details of a visual representation. The context for the whole project will revolve around categorical data in visual presentations. This means that different hues will be used to represent different categories of the data. The application of the automated color selection algorithms generated will help select optimum color choices and properties for maximum visibility. However, the basis for this representation remains the objective of the study. The areas that are going to benefit from this project range from graphical design, video production, cartography, printing companies, research, advertising, programming and many other sectors. 1.2 Statement of the problem The project tries to formulate a way of making the best choices as far as colors are concerned. The project is a research based algorithm to help select the colors and combine them then relate them to the real world variables in the visual presentation. When making visual presentations, there is often a problem of making color combinations to represent a state/scenario. This study provides a basis for calculating color difference, characteristics and comparing them with the real life situations to help make best color combinations for different situations. Designers are often faced with a problem of making calculated decisions about color schemes in their visual presentations. A tactical interplay with color schemes could help make a visual representation more presentable without necessarily changing the hue. The algorithm enhances the visibility of small characteristics of a dense visual representation and clear observation of important aspects of a visual representation. In most situations, designers find it hard to make small variables in their work visible; sometimes they don’t even notice the small variables themselves. This project helps designers to make clear contrast in their visual representations so that even small variables are vivid and clear. The other problem addressed in this work is the creation of categorical differences in the mind of the reader. Designers try to show important categorical differences in their visual presentations for the reader to understand and differentiate different variables in the work. The problem arises in setting up such differences in terms of color and color aspects. Some colors are close and when used together, the reader might not differentiate them. This makes a visual representation poor in quality. 1.3 Objectives of the study This study’s main objective is to generate a model for enhancing visibility of small features of a visual presentation. This helps designers to communicate required ideas to the reader effectively through color manipulation. To achieve this objective, we use the LCH LAB to calculate the color differences and the point saliency of different data points in the visual presentations. The project will help develop an algorithm for selecting optimum visibility of all categorical data in a visual representation. This will mainly help in making small features of a visual representation more visible, noticeable and clear. This will be achieved through application of the color difference formula, RGB color space, the LCH and XYZ. The project will generate a basis of calculating color aspects like difference, hue, lightness and saliency. This will be in relation to the color application in the visual presentation. Other factors will be considered like context and objectives of the work. The tool used for this task will be the LCH color lab. 1.4 Contribution This study tries to ease the art of graphics through generating a standard model of making choices. It creates a standard of making color combinations based on the sample data collected in the study. This study helps establish better methods of communicating messages through color from the designer to the analyst. The study helps advance the use and applications of the CIE LCH (lightness, chroma and hue) color space in conjunction with the RGB color space. RGB (red, green and blue) is normally used in the computer monitors. Computer graphics use this color space. This study helps students to apply more skills to the RGB color space to make their visual presentations smarter and more presentable. The approach used in the study combines various aspects of visual presentations to influence quality presentation of visual information. This study will help evolve the graphical applications in various industries to achieve higher values and heights. The use of recent developments in I.T and chromatics will be promoted by this study for quality performance of designers. This study will help focus the graphic designers on the various visual presentation aspects like point saliency, surround difference, visual search and visual attention. This will help improve quality of their works through fast, reliable and cost effective modeling of visual presentations. This study will also help them play with colors and establish different insights created by different combinations of colors. This will in turn generate broader experience in the use of color and color spaces in algorithms to make visual presentations smarter and more meaningful. The study also creates interest in the area thus; adding weight to the topic. Other researchers have room to improve on the findings and generation of other algorithms to fit different needs and requirements of visual presentations. This could also be borrowed by other studies like data analysis and presentation in statistics. The calculations in the study will also help the researchers to make their visual presentations more influential through the use of automatic color selection algorithm to communicate ideas. This study will influence and broaden the application of colors. It is planned to show the world how the creative use of color dimensions could be utilized to benefit the necessary works of human beings like data communication. Other scholars will be challenged to invent and innovate more uses of color dimensions and their properties or to criticize and develop the ideas created in this study. 2.0 Chapter two: literature review 2.1 Background and related work This study mainly focuses on automatic color selection techniques to help designers communicate desired ideas to the analysts through visual presentations. To achieve this objective, we consider the visibility of all categorical data in visual presentations through their color properties. This includes the hue, lightness and intensity. Studies have been done by many other scholars and researchers all aiming to maximize visibility of small features of visual presentations. The studies have established that saliency driven visibility is directly influenced by the lightness and the chroma of the colors used (Wang Y, Reibman A, Feng X. & Liu T. 2002). According to Rougier N. P. (2011), visual attention is deeply studied in terms of the brain. He explains the nature of the brain in terms of visual perception. His study proved that an object that has fine different characteristics from its neighborhood (environment) is easily visible and gains more attention from the eye than all other related data categories. He also points out that objects having similarities with their surround are difficult to locate visually. The object will take more time to locate hence it is not clearly visible. To solve this problem is a particular goal of this project. 2.2 Color Theory Color theory is the study of color schemes and the visual impacts to the analysts in visual presentations. In visual arts, color theory plays an important role since it helps the designers to understand the basic color concepts and approaches. Color theory mainly encompasses two dimensions. That is communicating color and applications of color. Communication using colors has been developed to the extent of using color labs to make color choices and/or palettes. This includes the Pantone color system, CIE LAB, CE LCH scale, TGB color scale and many others. This has facilitated the selection of hues to meet different objectives in visual presentations. Color can be used by designers for several reasons. One of them is to create categorical differences in visual presentations. This helps the designer to represent large volumes of data in a single visual display and yet maintaining visibility of even the smallest features. The other is to influence certain emotions and feelings in the mind of the reader. This is often achieved through variations in hue, chroma and lightness of all categorical data in a visual representation. For example, a designer could select a certain color scheme in his presentation to create harmony while another one represents a pixel with a specific hue to influence excitement. 2.3 RGB color space This term is derived from the three primary colors (red, green and blue). It is the basic colors used in computer monitors. All other colors are achieved through a technical and calculated mix of the colors. The use of this space helps ensure that designers do not have to quality through transformation of their work to other color spaces for printing. It also helps creation of a wider human-gamut. 2.4 CIE Lab color space It is normally referred to as lab due to its three important considerations. They are lightness, a* and b*. The vertical axis of the color space represents the lightness denoted by L*. Values are allocated ranging from zero to a hundred. The other two axes are the a* and b* that meet at the centre with the L*. They all coincide at right angles. a* has green at one end and red at the other end while b* has blue at one end and yellow at the other. The a* with green is labeled as –a* while that with red is +a*. This shows that a color cannot be both green and red. The point of intersections is either black/white or grey. In practice, the values in this color space are labeled from -128 to +127. This gives room for the fact that there are no extreme values in reality. This color space enables designers to select different colors at the specified lightness based on the (-128 to +127) scale. This color space has dimensions labeled b* and a*. These represent the hue factor. They are normally referred to as chromatic co-ordinates. These co-ordinates rotate around the color prism changing with each unit of the Euclidean distance. The possible colors resulted are wide and varied most of which are beyond the gamut of the human vision. 2.5 CIE LCH color space The lab constitutes of three dimensional color spaces that help combine the three dimensions of color. This helps the designer to manipulate a wide variety of colors for his visual presentation. The CIE LCH color model constitutes of a prism with three axes. These are hue, lightness and saturation (chroma). These three axes combine their variations to the three colors (red, green and blue) in the prism. This lab is device independent and goes to the extent of specifying the lighting conditions. This LAB is perceptually uniform. That is, a change in the color amount will result in a proportional change in the color perception in the human eye. The first axis L* runs from the top to the bottom of the prism. The top most part is white with a value of 100. This value reduces downwards as the lightness reduces. At the centre, we have color grey (value 50), that way up to zero (color black for zero lightness). The second axis is the chroma denoted C*. It is basically the radius of the prism. At the centre (zero radius), we have every possible grey of all the hues. This saturation is directly proportional to the radius. This means, at the maximum/full radius of the prism, we have the maximum chroma. This is equivalent to the hue of the color. The third axis is the hue denoted H*. A cross-section of the prism is considered i.e. the circle formed when the prism is cut into two equal halves. The outer edge is observed to have all possible hues. This runs all around the circle and is represented with the degrees. From 0-90 we have yellow, 90-180 green, 180-270 blue, 270-360/0 violet. This color scale is important in editing visual presentations through varying the angles, the lightness scale and the radius to achieve required visibility and saliency. A visual presentation could be perfected in terms of visibility through varying the lightness scale and the radius without necessarily varying the hue. This is the main use of this color space in the algorithm of this project. 2.6 Visual search and visual attention In a visual presentation, human beings will often search for specific categorical data. Eventually, there are data units that attract the human visual attention. This is all manifested by color and its dimensions. Any data category in a visual presentation that can be easily located without much effort is said to be visible. The visual perception of any data in a visual presentation is affected by these two. Designers will always try to enhance visual attention of small properties of a presentation through the use of high contrasts and saliency. The color difference formula is applied to select the optimal difference between colors in order to meet this objective. This study assumes that the saliency levels are directly proportional to the visibility of different data categories. This project tries to focus on a visual presentation. Analysts will always look for required data in the presentations. To achieve this objective, the data should be visible. The designers try to solve this problem by making color interplay of hue, saturation and value. For each data category to have maximum visual attention and be easily searchable, it must have a point saliency. This is the measure of how much a data category stands out from the rest of the data surrounding it. This is in terms of a data point color. It is also referred to as the measure of how much a class is visible from its centre surround. This project generates a point saliency formula that uses the color difference formula together with the mean color of the surrounding colors to rate the saliency level. A point with a high saliency level does not have similarities with its surrounding points nor its background. This makes it outstanding. Its saliency is determined by the surround within the surround radius. The computation of this measure is mainly based on the color differences with the surrounding. It also depends on the center point. The surrounds value is the average value of all the points surrounding the saliency point in context. In this project, we assume that a point/center is a single pixel denoted by p and the surround to be θp. The difference between the p and the θp is measured in single scale. It is calculated as the difference between the color of the pixel and the mean color of the surround. For a pixel p Є D where D is the data display space, the mean single scale difference Sp between its centre and the mean of the surround is; Sp = △E (Cp , NΘ-1 ∑ q∈Θp Cq ). where E(…) denotes the conceptual range of distance between the centre color and the surround color. CP is the color of the centre while cq is the associated mean color of the surround within the surround radius. Saliency levels are actually a measure of visibility of a data point in a visual presentation in relation to the surround and the background. Point saliency will be a tool to measure the visibility of a small data category in a visual presentation. The visibility measure will be a uniform metric that considers the capability of the analyst to view a class in a local region without much effort or wastage of time. The saliency measure will be applied in the algorithm. For any point (pixel) to be clearly visible, the number of similar points within its surround is highly considered. Equal points will be said to belong to the same class. High number of points of the same class in the same local region will denote high saliency hence, a high visibility. 2.7 color selection in visualization Color selection often involves color palettes. They are often used to show different categorical data types in different colors. Analysts will often view and rate data in a visual representation according to the saliency of the different categories. Effective color palettes should have large color differences and high saliency measures. This enhances the timely observation of key features by the analysts. In this project, the saliency levels are to be optimized for easy observation of even the smallest features of a visual presentation. We are also going to consider the factors of spatial context and coherence of the pixels for effective color palettes. Let us take two main factors affecting color palettes. These are order and separation. The order in which colors are chosen especially the front-background aspect, lightness and the algorithmic choices will greatly influence the effectiveness of the color palette used in a visual presentation. According to Stone M. (1991), expert color choices will normally depend on the analogy and contrast. He shows demonstrations that broaden the understanding of the two factors. Figure 2, of his writing shows that analogous colors are not salient (if two colors are similar, they are said to be analogous). In our context, we use color difference as the analogy factor. The contrast factor is shown by tactical separation of his colors. He proved that a high contrast between two data points (pixels) makes the more saturated one more salient. This can be seen through observation. An effective color palette should be presented with different levels of lightness for different types of categorical data. This enhances saliency and hence visibility of the data. This project’s algorithm will provide a useful tool in this task. Figure 6, of Stone M. (1991), the luminance of a color palette is highly determinant to the visibility of such a palette. He efficiently demonstrates how luminance alone affects the visibility of categorical data (pixels in our case). In this project, we shall refer this as lightness since we are using the CIE LCH color space. He also explains that effective color palettes should not over depend on the hue variation i.e. using too many hues in a visual presentation. He showed that color palettes are most presentable with few colors and some variations in lightness and saliency. Furthermore, use of many colors especially using computer visual displays may go beyond the human gamut hence limit the visibility of small data categories. REFERENCES Chang Y, Lee d. & Archibald J. K.: Using color variation to mimic human visual perception for color image quantization retrieved from http://www.et.byu.edu/~jka/papers/quantpaper.pdf Color therapy healing: color properties retrieved from http://www.colourtherapyhealing.com/colour/colour_properties.php Glencoe (2011): three properties of color retrieved from http://www.glencoe.com/sites/common_assets/art/artquests/color_properties.htm Kaya N. & Epps H. H. (September 2004): relationship between color and emotion: a study of college students retrieved from http://findarticles.com/p/articles/mi_m0FCR/is_3_38/ai_n6249223/ Learningspeed.com (2006): color in everyday life retrieved on fifth November 2011 at http://www.learningseed.com/_guides/1150_Color_In_Everyday_Life_Guide.pdf Morton J. L. (2011): welcome to color matters retrieved from http://www.colormatters.com/ Rougier N. P. (2011): Visual attention Retrieved from http://www.loria.fr/~rougier/research/attention.html Stone M. (1991): expert color choices for presenting data Retrieved from http://www.stonesc.com/pubs/Expert%20Color%20Choices.pdf Voice of America (07th February 2009): study finds color boosts brain performance retrieved from http://www.voanews.com/english/news/a-13-2009-02-07-voa1-68811667.html Weijer J. (2005): color features and local structure in image retrieved from http://lear.inrialpes.fr/people/vandeweijer/papers/thesis.pdf Williams N, C, N & Y. (2002): color perception is mediated by a plastic neural mechanism that is adjustable in adults. Neuron, 15; 35(4) pp. 783-792 Wang Y, Reibman A, Feng X. & Liu T. (2002): Saliency inspired modeling of packet-loss visibility in decoded videos. Retrieved from http://enpub.fulton.asu.edu/resp/vpqm/vpqm09/Proceedings_VPQM09/Papers/VPQM091TaoLiu.pdf List of abbreviations RGB – red, green and blue CIE LCH LAB –international commission of illumination (lightness, chroma and hue laboratory). Read More

Designers are often faced with a problem of making calculated decisions about color schemes in their visual presentations. A tactical interplay with color schemes could help make a visual representation more presentable without necessarily changing the hue. The algorithm enhances the visibility of small characteristics of a dense visual representation and clear observation of important aspects of a visual representation. In most situations, designers find it hard to make small variables in their work visible; sometimes they don’t even notice the small variables themselves.

This project helps designers to make clear contrast in their visual representations so that even small variables are vivid and clear. The other problem addressed in this work is the creation of categorical differences in the mind of the reader. Designers try to show important categorical differences in their visual presentations for the reader to understand and differentiate different variables in the work. The problem arises in setting up such differences in terms of color and color aspects.

Some colors are close and when used together, the reader might not differentiate them. This makes a visual representation poor in quality. 1.3 Objectives of the study This study’s main objective is to generate a model for enhancing visibility of small features of a visual presentation. This helps designers to communicate required ideas to the reader effectively through color manipulation. To achieve this objective, we use the LCH LAB to calculate the color differences and the point saliency of different data points in the visual presentations.

The project will help develop an algorithm for selecting optimum visibility of all categorical data in a visual representation. This will mainly help in making small features of a visual representation more visible, noticeable and clear. This will be achieved through application of the color difference formula, RGB color space, the LCH and XYZ. The project will generate a basis of calculating color aspects like difference, hue, lightness and saliency. This will be in relation to the color application in the visual presentation.

Other factors will be considered like context and objectives of the work. The tool used for this task will be the LCH color lab. 1.4 Contribution This study tries to ease the art of graphics through generating a standard model of making choices. It creates a standard of making color combinations based on the sample data collected in the study. This study helps establish better methods of communicating messages through color from the designer to the analyst. The study helps advance the use and applications of the CIE LCH (lightness, chroma and hue) color space in conjunction with the RGB color space.

RGB (red, green and blue) is normally used in the computer monitors. Computer graphics use this color space. This study helps students to apply more skills to the RGB color space to make their visual presentations smarter and more presentable. The approach used in the study combines various aspects of visual presentations to influence quality presentation of visual information. This study will help evolve the graphical applications in various industries to achieve higher values and heights. The use of recent developments in I.

T and chromatics will be promoted by this study for quality performance of designers. This study will help focus the graphic designers on the various visual presentation aspects like point saliency, surround difference, visual search and visual attention. This will help improve quality of their works through fast, reliable and cost effective modeling of visual presentations. This study will also help them play with colors and establish different insights created by different combinations of colors.

This will in turn generate broader experience in the use of color and color spaces in algorithms to make visual presentations smarter and more meaningful. The study also creates interest in the area thus; adding weight to the topic.

Read More