Scientific Illustration → Part 1: Traditional Techniques and NPR Approaches Mario Costa Sousa Department of Computer Science University of Calgary 2500 University Drive N.W. Calgary, AB, Canada T2N 1N4 Email: [email protected]

1

What is Scientific Illustration?

Scientific illustrations are drawings of measured accuracy, depicting subtleties without ambiguities. It is an accurate work that is also pleasing to the eye in terms of balance and artistic handling of the subject. Two quotes express this very well: “. . .aesthetic qualities will give to the illustration life and a measure of charm. These will help to put over the facts and to fix them more firmly in the readers memories.” Geoffrey Lapage [28] “If the technical illustrator will just keep in mind that the product and its components have already been designed and his main job is to present it accurately and with emphasis as required he will have easy sailing.” Anthony D. Pyeatt [42] Drawing or Photo . . . Which? It is important to notice that scientific illustrations might or might not be visually realistic. Though often highly representational (”realistic-looking”), the main purpose of a scientific illustration is to communicate information and not necessarily to look real. This is an important aspect that makes scientific illustrations differ from photo-realism and other representational genres. For example, in medical subjects, there are four instances where good illustration is the best (and possibly the only) medium to use; this is the case where: (1) Areas of reference exist physiologically but not gross anatomically; (2) Superimposing one structure upon another gives related information; (3) Section views show instruments in place in body cavities, etc; (4) Eliminating much visual garbage from a photo can produce a simpler explanation. There are also four instances where good photographs are better than drawings and cost far less; this is the case where (1) Overall posture, or before and after pictures are necessary; (2) Vast areas such as large skin areas, burns, etc., are to be shown; (3) Emotional impact is achieved only by authentic photos of, say, medico-legal problems such as gunshot wounds, etc. (4) A multitude of detail is necessary as in retinal pathology or photomicrographic studies. The recorded history of scientific illustration dates back in the 1200s up to the present [34]. Many resources are available, including societies, journals, conferences, and of course excellent books [13, 17, 20, 21, 25, 29, 38, 45, 56, 59]. Today, various opportunities exist for the professional scientific illustrator and for anyone seeking formal education on the field. The use of nonphotorealistic rendering (NPR) for scientific illustration is gaining momentum in the research community with very interesting results and exciting future directions. This first part of the course notes is designed to be a basic reference covering the fundamental topics of traditional scientific illustration and existing NPR research: production pipeline (sec. 2),

NPR system framework (sec. 3), rendering strategies (sec. 4), media and rendering techniques (sec. 5) and subjects (sec. 6). There are, of course, many other techniques besides the ones outlined that use media of various types for producing scientific illustrations, either alone or in combination, including water-colored pencil, carbondust, gouache and acrylics, airbrush, among others.

2

Communication Pipeline

Scientific illustration is produced for a specific kind of visual communication in the sciences. The illustrator must therefore be aware of the viewers level of knowledge and must relate the message in a logical sequence providing him/her with a balanced amount of information. The subject must be rendered with scientific accuracy and artistic integrity. Table 1 shows the responsibilities and communication pipeline between the scientist and illustrator. This table also indicates where potential NPR tools fit in to help scientific illustration with their work. SCIENTIST Provides material description specimen Checks detailed preliminary drawing Checks corrections Checks rendering Checks labeling -

ILLUSTRATOR Requests information Records information Studies specimen Makes rough drawing Makes detailed preliminary drawing Corrects preliminary drawing Produces rendering Labels drawing Return specimen

Table 1: Traditional communication pipeline for scientific illustration production.

3

NPR System Framework

General NPR systems can be broken down into four basic parts as shown in Table 2. Notice that each part builds upon the others and is essential to developing the framework for higher level NPR methods and tools. We can adapt this system organization to a NPR framework for scientific illustration composition (fig. 1). The subject (sec. 6) is

COMPOSITION ↓ RENDERING ↓ PRIMITIVES ↓ MEDIA

stages, principles, ... domain (image, object, hardware), ... strokes, brushes, erasers, blenders, ... pen-and-ink, pencil, watercolor, ...

Table 2: Main components of general NPR systems.

ridges and valleys [49], creases [23], and lines related to the principal directions of curvature [14, 19, 47]. For more detail, refer to the two lectures from course component Interpreting Form. A recent approach involves the use of geomorphologic shape measure calculation schemes to provide a large and computationally stable collection of shape measures [53].

4.2

Lightning Conditions

In some situations, conventional lighting conditions make it difficult to see certain structural aspects of a specimen or conceptualization, and alternative lighting may provide a better view. Aside from the manipulation of the effects of light, several techniques are used by illustrators to achieve a specific pictorial objective. They include isolation of the subject and selective emphasis of detail. More details on the lecture Basic Principles of Composition. In NPR, Interrante [22] defines a technique for illustrating layered transparent surfaces, in three dimensions. Interrante is presenting more details about her approach in the lecture Shape Analysis. Hamel [15] develops an empirical lighting model based on observations on scientific and technical illustrations as well as other documented techniques of illustrators and artists. He also presents a model for the transmission of light through transparent surfaces, with results resembling the traditional ink line illustrations.

Figure 1: Scientific illustration composition.

5 5.1

first selected and studied. Selection is mainly related to the drawing composition principle of unity, where all the parts must be related, expressing one main thought (refer to the course notes on Drawing Composition for more details). To study the subject means to perform preliminary form interpretation (subsec. 4.1). The subject is then rendered in progressive stages. At each stage, principles of composition are applied (see course notes on drawing composition), and specific rendering strategies are adopted to cope with form interpretation (subsec. 4.1) and lighting adjustments (subsec. 4.2). Finally, primitives and media models are accessed. They are modeled after traditional techniques used in scientific illustration: ink lines (subsec. 5.1), scratch board (subsec. 5.2), pencil (subsec. 5.3), coquille board (subsec. 5.4), and wash (subsec. 5.5).

4 4.1

Rendering Strategies Interpreting Form

Form interpretation starts during subject studying (fig. 1, Subject) where the goal is to visually eliminate extraneous details and reduce the subject to most representative features. Shape characteristics (i.e. contours, folding regions, surfaces areas, volumes, curvatures) are accurately identified and measured. This preliminary measure is then used later on as a basis to render the forms of the subject (fig. 1, Rendering). Regions related to the shape measures are lightly outlined using pencils, and filled with strokes, with a gesture that conveys a careful constructed look [20, 51]. The overall composition of strokes to give form to the structures should begin at the center of interest and diminish in degree as one move away from this focal point. Peripheral areas are left open, or only slightly rendered, is a relief to the eye, and can enhance and direct attention to important content (refer to Draing Composition lecture notes). In NPR, most of the research has been focused on silhouettes extraction. Other types of lines extracted from the model included

Traditional Illustration Techniques Line Drawings (Ink)

A line drawing is termed as any illustration composed entirely of sharp, distinct lines and dots of solid black India ink applied with flexible nib pens on white paper. Pen and ink is the most useful and challenging drawing technique used in scientific illustration. It can depict any subject except those few requiring exact representation of surface texture or color. Its reproduction is more reliable and economical than that of continuous tone. Outline Approaches to final rendering of a pen and ink drawing vary greatly among illustrators, but it is generally best to approach the outline first. One should use a sensitive line, varying its weight (or thickness) to delineate form without reliance on rendering (subsec. 4.1). These variations in line weight can accentuate important points and add depth and activity to the drawing. Once forms are defined, one can still use outline for rendering but this is difficult to achieve because the line must suggest texture and limits of areas and nuances of light on form, and still dramatize the point to be made by the illustration. Line Weight One important consideration that applies to any line drawing technique at any drawing stage (i.e. initial sketch, finished rendering) is the weight (or thickness) control of the lines. As mentioned above, proper line weighting results in drawings with good design and balance. Some examples of line weighting strategies focus on depicting various nuances such as of lights and darks, contours and internal structures within the form, types of materials (i.e. hard and soft), line junctions, depth perception, and transparency. Existing NPR approaches control the weight of strokes based on depth [24, 33, 37], on tone [24], and more recently on surface curvature [53, 54].

Figure 2: Examples of stippling.

Crosshatching Crosshatching consists in crossing a series of lines of various lengths, widths and at various angles with which the artist constructs areas of tone and texture. This is a much common rendering technique. It can be a quick and effective for drawings intended to be so reduced that the crosshatching suggests modeling of form and not the specimens actual texture. Notice the egg and shadow in Figure 3, both indicated with simple hatching weighted strokes. Figure 4 shows a more detailed crosshatched botanical illustration, with strokes following curvature of the petals. In NPR crosshatching reproduction has been explored in various domains with algorithms and tools still being developed. Refer to the course notes on Indexed Taxonomies for specific NPR references related to hatching/crosshatching (under ink/lines category).

Figure 3: In this illustration, the texture of feathers is suggested by using a combination of short-lines and no outlines. Hatched weighted lines are applied to selected parts of the bird, egg, and shadows.

Short-lines and Stippling Short, straight lines are also very common in scientific illustration, allowing for some crosshatching and also the simulation of a great variety of textures at different levels of precision (notice the birds in Figures 3 and 5). Stippling is the effect obtained by using a series of properly scaled and spaced dots. It is the most precise of all pen techniques.

Figure 4: Precise crosshatching of botanical illustrations.

Figure 5: Precise illustrations of flower and bird using scratch boards. Notice the key effects of using scratch board on the leftmost sample.

Figure 2 shows examples of stippling for two archeological artifacts and for a texture pattern with a suggestion of light and shade. Notice that lines and dots should be clearly separated from one another, regardless of the technique used; Shading effects may be achieved by gradually decreasing the thickness of lines or dots, or by decreasing the distance between them. In NPR, different approaches for placing small pen-and-ink primitives over 3D models have been proposed. Winkenbach and Salesin [57] introduce “stroke textures” allowing procedural accumulation of strokes for stippling and other textures made with small lines such as “grass”. Elber [11] presents a technique for spreading small strokes uniformly across freeform surfaces. Deussen and Strothotte [9], use a particle-based distribution to render clusters of leaves as small line primitives. Building on the previous technology of stroke textures [57], Praun et al. [41] introduce “tonal art maps” which organizes pre-rendered strokes as a sequence of mipmapped images. Research on stippling has focused on the geometric relation between the stippled dots [8] and on interactive direct volume illustration systems [31] (more details on the second part of this course). Secord [50] introduces a fast probabilistic method that places small arbitrarily-shaped primitives, including stippling. More recently, Sousa et al. [53] presents an approach that distributes pen marks on a fixed basis across feature regions of dense meshes (i.e. one stroke per mesh edge). Also, Pastor et al. [39] presents a method that directly distributes and controls the density of stippled marks over 3D meshes.

5.2

Scratch Board

The most attractive line drawings are done on scratch board, which is a heavy sheet of paper with a fine coating of soft ink substance giving it a smooth, even finish on one side. A Scratchboard drawing is made using a sharp-pointed instrument such as an etching point or exacto-type blade to make white lines in the inked areas by scratching, scraping or scuffing into the inked surface to remove ink. The resulting line is very white and clean. Figure 5 shows examples of scientific illustration using scratch boards. In NPR, scratch board can be related to existing work on halftoning and engraving. Refer to the course notes on Indexed Taxonomies for the specific references.

5.3

Pencil Drawings

Graphite Pencil: there are three elemental ingredients in graphite pencil drawings:

Figure 6: Graphite pencil drawing of an insect (left) and smudging it together with eraser to suggest surface smoothness (right).

1. Solid lines of varying weight is of special importance in scientific illustrations. Outlines should be light and may be erased after the drawing is completed, where the internal details may be blocked out by shading. 2. Shading with uniform or gradually weakening tones. 3. Smudging: After shading is finished, a fine pointed smudger (Stump) is used to smear the graphite from the dark area toward the lighter area, using a light pressure. Very slight smudging is required in the lightest area, leaving some of the white area white. More graphite should be used for darker parts. For an irregular surface appearance, a soft- pointed eraser should be lightly used. Colored Pencils: are also used for scientific illustration. If handled properly, they can give textural and color effects very difficult to achieve with any of the conventional wet color media. Color should be applied in most cases with the point of the pencil as crosshatching, side by side, at random, or in combination. For shading with colored pencils, begin with middle tones, working toward the darker and lighter values. Use combination of colors to build up brilliant

effects, such as yellows, yellow-greens, and blues to make a bright green. Lighten when necessary with white pencil on top of the other colors. In general, apply lighter colors over darker ones. Do not smudge or blend the strokes with a stump because this will produce a shiny surface that will no longer accept pencil. Figure 6 shows the result of using graphite pencil and smudging. Refer to the course notes on Indexed Taxonomies for specific NPR references related to pencil.

Figure 8: Watercolor rendering for botanical illustration. Notice the different level of detail and emphasis used in the composition.

Refer to the course notes on Indexed Taxonomies for specific NPR references related to wash/watercolor.

6

Figure 7: Different patterns of coarseness and structure of coquille boards (left). An adze artifact rendered using coquille board (right); notice the resemblance to stippling.

5.4

Coquille Board

This technique, although restrictive, is probably the fastest method. The coquille board is a special paper with a rough surface, and different patterns of coarseness and structure (fig. 7, left). Outlines are drawn first with pencil, and then blackened with India Ink. Shading is accomplished through the use of a wax pencil. This technique saves a great deal of time and enables the production of fine artwork. The density of dots may not be regulated, as by stippling, but their size may be regulated by using heavier pressure on the wax pencil (fig. 7, right). In NPR, existing models for pencil/charcoal paper interaction can be used to reproduce coquille board effects [2, 10, 32, 52].

5.5

Watercolor and Wash

This technique has long been used in scientific illustration, in particular for medical subjects. Under this category fall illustrations made with brush and any water soluble material. The term watercolor implies full color, and the term wash implies black and white (as well as a way to apply paint), but the media for both are applied similarly. There are two basic ways of applying wash (regardless of color) to a surface: 1. Wet-on-wet: ensures good control, even tones, smooth gradation, and a softer overall drawing; 2. Wet-on-dry: offers a high degree of control and the ability to render fine detail and pattern meticulously.

Subject Matter

This section presents the main classes of subjects of scientific illustrations: anatomy [16, 18, 30, 35, 46, 58], botany [3, 26, 44, 55], zoology [4, 6, 7, 12, 27] and archeology [1, 5, 36, 40, 48]. Almost any media and technique can be used to render those subjects. Pen-and-ink is most often used because it is cheapest and easiest to reproduce well. Most importantly, pen-and-ink can represent virtually any shape if used properly [20, 43, 51]. In some fields, there are traditions of emphasizing certain techniques, and there are even conventions encouraging a certain appearance for a drawing, but both illustrator and scientist should keep an open mind about the way a subject might be best depicted. As we pointed out, scientific illustrations require proper measurement and exact depiction of the subject. However, medical illustration does not always require that; often it is more a conceptual than a literal depiction of the subject, thus allowing more latitude for artistic freedom. Similarly, zoological and botanical illustrations are not always bound by the requirements of exact representation. They have more of the freedom of the fine artist and are able to work in color more extensively, whereas other scientific illustrators use color less often because of cost restrictions.

References [1] L. R. Addington. Lithic Illustration: Drawing Flaked Stone Artifacts for Publication (Prehistoric Archaeology and Ecology Series). University of Chicago Press, 1986. [2] T. W. Bleser, J. L. SIbert, and J. P. McGee. Charcoal sketching: Returning control to the artist. ACM Transactions on Graphics, 7(1):76– 81, January 1988. [3] W. Blunt. The Art of Botanical Illustration: An Illustrated History. Dover Pubns, 1994. [4] J.E. Bohlke and C.G. Chaplin. Fishes of the Bahamas and Adjacent Tropical Waters. Livingston Press, 1968. [5] I.W. Cornwall. Bones for the Archeologist. Phoenix House, London, 1956.

[6] G.B. Cuvier. Cuvier’s Animals: 867 Illustrations from the Classic Nineteenth-Century Work. Dover Publications, 1996.

[32] A. Majumder and M. Gopi. Hardware accelerated real time charcoal rendering. In Proc. of NPAR ’02, pages 59–66, 2002.

[7] S.P. Dance. The Art of Natural History Animal Illustrators and Their Work. Overlook Press, 1978.

[33] L. Markosian, B. J. Meier, M. A. Kowalski, L. S. Holden, J. D. Northrup, and J. F. Hughes. Art-based rendering with continuous levels of detail. In Proc. of NPAR ’00, pages 59–66, 2000.

[8] O. Deussen, S. Hiller, C. van Overveld, and T. Strothotte. Floating points: A method for computing stipple drawings. Computer Graphics Forum (Proc. of EuroGraphics ’00), pages 40–51, 2000.

[34] G. Mikel. Art History: Chronology of Scientific Illustration. http://www.scientificillustrator.com/artists/history.html.

[9] O. Deussen and T. Strothotte. Computer-generated pen-and-ink illustration of trees. In Proc. of SIGGRAPH ’00, pages 13–18, 2000.

[35] F. H. Netter and J. T. Hansen. Atlas of Human Anatomy. Novartis Medical Education; 3rd edition, 2003.

[10] F. Durand, V. Ostromoukhov, M. Miller, F. Duranleau, and J. Dorsey. Decoupling strokes and high-level attributes for interactive traditional drawing. In Proc. of Eurographics Workshop on Rendering ’01, pages 71–82, 2001.

[36] E.W. Nield. Drawing and Understanding Fossils: A Theoretical and Practical Guide for Beginners, With Self-Assessment. Pergamon Press, 1987.

[11] G. Elber. Interactive line art rendering of freeform surfaces. Computer Graphics Forum (Proc. of EuroGraphics ’99), 18(3):1–12, 1999. [12] W. Ellenberger. An Atlas of Animal Anatomy for Artists. Dover Publications, 1949. [13] B.J. Ford. Images of Science: a History of Scientific Illustration. Oxford University Press, 1999. [14] A. Girshick, V. Interrante, S. Haker, and T. S. Lemone. Line direction matters: An argument for the use of principal directions in 3d line drawings. In Proc. of NPAR ’00, pages 43–52, 2000. [15] J. Hamel. Alternative Lighting Methods for Computer Generated Line Drawings. PhD thesis, University of Magdeburg, 2000. [16] J. V. Hansen, S. Porter, and M. Kemp. The Physician’s Art: Representations of Art and Medicine. Duke Univ Pr (Trd), 2000.

[37] J.D. Northrup and L. Markosian. Artistic silhouettes: A hybrid approach. In Proc. of NPAR ’00, pages 31–37, 2000. [38] C.S. Papp. Technical Illustration. Higgins Ink Co., 1960. [39] O.E.M. Pastor, B. Freudenberg, and T. Strotthote. Animated, real-time stippling. IEEE Computer Graphics and Applications (Special Issue on Non-Photorealistic Rendering), to appear, 2003. [40] S. Piggott. Antiquity Depicted Aspects of Archeological Illustration. Thames and Hudson, Np, 1978. [41] E. Praun, H. Hoppe, M. Webb, and A. Finkelstein. Real-time hatching. In Proc. of SIGGRAPH ’01, pages 579–584, 2001. [42] A.D. Pyeatt. Scientific Illustration: Theory and Practice. WM.C. Brown Company Publishers, 1968. [43] P. Rawson. Drawing. University of Pennsylvania Press, 1987. [44] M. Rix. The Art of the Botanist. Guildford: Lutterworth Press, 1981.

[17] J. G. Heck. Heck’s Pictorial Archive of Nature and Science. Dover Publications, 1994.

[45] H. Robin and D.J. Kevles. The Scientific Image: From Cave to Computer. W H Freeman Co., 1993.

[18] R. Herrlinger. History of Medical Illustration From Antiquity to a. D. 1600. Editions Medicina Rare Ltd. , New York, 1970.

[46] J. W. Rohen, C. Yokochi, and E. Lutjen-Drecoll. Color Atlas of Anatomy: A Photographic Study of the Human Body. Lippincott, Williams Wilkins; 4th edition, 1998.

[19] A. Hertzmann and D. Zorin. Illustrating smooth surfaces. In Proc. of SIGGRAPH ’00, pages 517–526, 2000. [20] E.R.S. Hodges. The Guild Handbook of Scientific Illustration. Van Nostrand Reinhold Company, 1989. [21] F. Howard. The Science of Drawing. Being a Progressive Series of the Characteristic Forms of Nature. William Pickering, London, 1840. [22] V. Interrante. Illustrating Transparency: communicating the 3D shape of layered transparent surfaces via texture. PhD thesis, University of North Carolina at Chapel Hill, 1996. [23] R. D. Kalnins, L. Markosian, B. J. Meier, M. A. Kowalski, J. C. Lee, P. L. Davidson, M. Webb, J. F. Hughes, and A. Finkelstein. WYSIWYG NPR: Drawing strokes directly on 3d models. ACM Transactions on Graphics (Proc. of SIGGRAPH ’02), pages 755–762, 2002. [24] M. Kaplan, B. Gooch, and E. Cohen. Interactive artistic rendering. In Proc. of NPAR ’00, page 6774, 2000. [25] M. Kemp. Visualizations: The Nature Book of Art and Science. University of California Press, 2001.

[47] C. Rossl, L. Kobbelt, and H.P. Seidel. Line art rendering of triangulated surfaces using discrete lines of curvature. In Proc. of WSCG ’00, pages 168–175, 2000. [48] M. J. S. Rudwick. Scenes from Deep Time: Early Pictorial Representations of the Prehistoric World. University of Chicago Press, 1995. [49] T. T. Sasada. Drawing natural scenery by computer graphics. Computer-Aided Design, 19(4):212–218, May 1987. [50] A. Secord. Random marks on paper: Non-photorealistic rendering with small primitives. Master’s thesis, University of British Columbia, 2002. [51] G. Simmons. The Technical Pen. Watson-Guptill Publications, 1992. [52] M.C. Sousa and J.W. Buchanan. Observational model of graphite pencil materials. Computer Graphics Forum, 19(1):27–49, 2000. [53] M.C. Sousa, K. Foster, B. Wyvill, and F. Samavati. Precise ink drawing of 3D models. Computer Graphics Forum (Proc. of EuroGraphics ’03), to appear, 2003.

[26] R. King. Botanical Illustration. Wraps. C. N. Potter: distributed by Crown Publishers, New York, 1979.

[54] M.C. Sousa and P. Prusinkiewicz. A few good lines: Suggestive drawing of 3D models. Computer Graphics Forum (Proc. of EuroGraphics ’03), to appear, September 2003.

[27] D.M. Knight. Zoological Illustration: An Essay Towards a History of Printed Zoological Pictures. Shoe String Press, 1977.

[55] K.R. West. How to Draw Plants: The Techniques of Botanical Illustration. Timber Pr, 1996.

[28] G. Lapage. Making science readable. Medical and Biological Illustration, 12(2):114–120, 1962.

[56] J. J. White and G. D. R. Bridson. Plant, Animal and Anatomical Illustration in Art and Science: A Bibliographical Guide from the 16th Century to the Present Day. St Pauls Bibliographies, 1990.

[29] J.B. Lee and M. Mandelbaum. Seeing Is Believing: 700 Years of Scientific and Medical Illustration. New York Public Library, 1999. [30] W. E. Loechel. Medical Illustration: A Guide for the Doctor-Author and Exhibitor. Charles C Thomas Publisher, LTD., 1964. [31] A. Lu, C. Morris, D. Ebert, P. Rheingans, and C. Hansen. Nonphotorealistic volume rendering using stippling techniques. In Proc. of IEEE Visualization ’02, pages 211–218, 2002.

[57] G. Winkenbach and D. H. DavidH. Salesin. Computer-generated penand-ink illustration. In Proc. of SIGGRAPH ’94, pages 91–100, 1994. [58] E. Wolff. Anatomy for Artists. H.K. Lewis Co. Ltd., 1958. [59] P. Wood. Scientific Illustration: A Guide to Biological, Zoological, and Medical Rendering Techniques, Design, Printing, and Display. John Wiley and Sons, 2nd edition, 1994.