Glossar-English:Image Morphology

Sub-Item to: Image Syntax

Original: Bildmorphologie

---

Placement of image morphology In general, the term ‘image morphology’ is not used with any specific meaning distinct from ‘image grammar’ or ‘image composition’. All of these terms refer to an analytical view of the image carrier as composed of parts relevant to the image function and essentially determined by visually perceptible properties, which can also be used in other image carriers that differ in how the parts are put together. It is the assembly of the parts into a whole that, together with other (non-syntactic) factors, determines the (type) identity of the image carrier and, thus, ultimately determines the possible uses of the image carrier as an image. However, the distinction between grammatical and morphological aspects in the narrower sense when considering language (⊳ morphology and syntax) suggests that the expression ‘image morphology’ should be given a more specific meaning and thus set it apart from image grammar (which thus becomes image grammar in the narrower sense). Image grammar in the narrower sense tries above all to prove syntactic compositionality in pictures in the sense of the formal (Chomsky) grammars, which presuppose the distinction between​ »sentence«​ and​ »word«;​ Through a limited set of replacement rules over a finite set of sentence constituents as intermediate stages (“nonterminal symbols”, such as ‘noun phrase’), they allow for infinitely many sentences to be derived or analyzed from a finite number of words (in the mental lexicon) in a unique way ([Chomsky 1957a]Chomsky, Noam (1957): Syntactic Structures. Mouton, Den Haag. ; [Sachs-Hombach 1999a]Sachs-Hombach, Klaus (1999): Gibt es ein Bildalphabet?. In: Sachs-Hombach, K. & Rehkämper, K. (Hg.), Bildgrammatik. Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg. ). On the other hand, image morphology in the sense used here is interested in syntactic image compositionality of a different kind: Can image carriers be described with the help of general grouping rules – for instance, analogous to the much “softer” Gestalt laws operating in the geometric continuum – as consisting of “pictorial primitives” that are not already considered signs (words) and originate from a possibly unlimited basic set, whereby the condition of uniqueness of derivation might also be weakened? In analogy to the word formation rules in extremely polysynthetic-​fusional languages, but without assuming that it makes sense to apply the distinction between​ »sentence«​ and​ »word«​ to pictorial sign systems, such a pictorial morphology would have to do justice to the characteristic property of the syntactic density of pictorial sign systems. If such a morphological pictorial compositionality can be rationally introduced, then images, although complex sign systems, are very clearly distinguished from language sign systems (⊳ Iconic Difference). They lack the subdivision of the individual overall sign actions into words: partially independent, albeit more or less (in Frege's sense) unsaturated partial sign actions that always need to be supplemented. In particular, whether the uses of isolated syntactic elements of pictorial sign systems are always already unsaturated sign actions remains open. The physical properties relevant to an object's function as a picture carrier can certainly be found in the visually perceptible geometric arrangement of colored areas above all. In this sense, the syntactic elements into which the image carrier is decomposed in an intrinsic value consideration of image syntax can be regarded as pictorial morphological elements. Beyond their intrinsic value, these are neither necessarily charged with a specific meaning – a specific depiction value – nor do they necessarily have a precisely defined pragmatic function – a fixed representational value. Visual Gestalts, coloremes, and pixemes Figure 1: As an example: Gloria Temarre Petyarre: «Arnkerrthe​(Mountain Devil Lizard)-​Dream» In short, exactly those entities into which the picture carrier – or more precisely, the part of its surface marked by the frame – appears divided in visual perception form the morphological repertoire of pictures. Psychologically, this division is determined by the Gestalt laws: they determine which spatial locations are seen as connected, not only in the sense of an undivided, coherent area of ​​the same color but also in the sense of higher-order groups, such as sequences of strokes of the same color. In Figure 1, for example, this leads to the fact that, in addition to the red, brown, black, yellow, and white elementary areas, the groups of yellow or white-framed, dark-filled arches and bars are also perceived as visual shapes that belong together. In her influential book on image syntax ([Saint-Martin 1990a]Saint-Martin, Fernande (1990): Semiotics of Visual Language. Indiana University Press, Bloomington, IN. ), Fernande Saint-​Martin introduces the so-called ‘coloreme’ as the basic morphological unit: [A coloreme] corresponds to that aggre­gate of visual vari­ables per­ceived in the visual repre­sen­tation by way of an ocu­lar fixa­tion, or focus of the gaze. … A col­oreme is de­fined […] as the zone of the visual linguis­tic field corre­lated to a centra­tion of the eye. It is consti­tuted by a mass of ener­getic matter present­ing a given set of visual vari­ables. ([Saint-Martin 1990a]Saint-Martin, Fernande (1990): Semiotics of Visual Language. Indiana University Press, Bloomington, IN. : p. 5). Figure 2: Visualization of a coloreme according to Saint-​Martin (schematic representation): The circular foveal centering area that determines the coloreme is highlighted and enlarged, while the rest is shown as somewhat darkened Saint-​Martin's understanding of the coloreme obviously focuses on momentary psychophysical aspects: At any given time, only one ocular fixation can take place, and consequently, only one coloreme can be perceived (see Fig. 2). However, a “colorematic (or coloremic) analysis” should build on this basis, which describes the transformations which a coloreme undergoes by its interrelations with the other coloremes of its immediate entourage through macular centrations. The analysis proceeds thus at a first regrouping of coloremes through the topological relations, which establish the first perceptual construction and structure the energetic exchanges between coloremes. ([Saint-Martin 1990a]Saint-Martin, Fernande (1990): Semiotics of Visual Language. Indiana University Press, Bloomington, IN. : p. 194). In their dynamics and direct dependence on the psychophysical idiosyncrasies of a perceiving individual, coloremes are primarily theoretical entities. In practice, Saint-​Martin suggests dividing the image area into a regular 5*5 grid, which serves as the basis for an approximate description of the possible or probable coloremes. Each grid is divided into a 5*5 sub-​grid, which picks up the division into foveal centers and macular peripheral areas (ibid.: p. 197ff). In order not to be too closely bound to Saint-​Martin's rather specific conception, it is advisable to more generally consider the visual shapes – ultimately related to a hypothetical standard viewer – in the context of image syntax in a rather structuralist manner and, in analogy to the linguistic expression ‘morpheme’, to designate them as ‘pixemes’. At the same time, in a first approximation, the dynamics, which makes the morphological description of a picture carrier considerably more difficult in Saint-​Martin, can also be disregarded. Pixeme attributes Saint-Martin distinguishes two types of properties of the syntactic-morphological elements of pictorial signs, which are often interpreted in the following way (cf., e.g., [Dölling 1999a]Dölling, Evelyn (1999): Kategorialstruktur ikonischer Sprachen und Syntax der visuellen Sprache. In: Sachs-Hombach, K. & Rehkämper, K (Hg.), Bildgrammatik: Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg. ): Plastic properties belong to the material of the image carrier, while other properties are in the eye of the beholder and are of a more visual, i.e., perception-dependent, nature. The geometric shapes and their topological relations are given as typical examples of the latter type of property, whereas colors and textures are examples of the properties of the material itself. Coloremes are always combinations of plastic and visual-perceptual properties. In fact, the more general concept of the pixeme can also be logically analyzed into a purely geometric base structure on the one hand and a conceptual field of marker dimensions that make these structures visible on the other because space as such would not be perceptible. Only the segmentation into areas that belong together – i.e., areas that are marked the same way – results in a structuring into the spatial elements of a whole. However, the distinction between geometry (i.e. spatiality) and color does not (or at least not essentially) depend on the differentiation between properties that belong to the material of the picture carrier – and should therefore be regarded as objective properties – and properties that the viewer constructs – and should therefore be valued as subjective properties. Rather, image philosophers can discuss colors and the relationships between them on the one hand, and spatial entities and the geometric or topological relationships between them on the other, without having to mix the two arguments. They can be viewed as independent of each other and treated as regulated by – at least at first glance – autonomous conceptual fields. Argumentations about pixemes, on the other hand, have to combine the logic of color and the logic of space, i.e., unite them in a conceptual synthesis. Fundamental properties of pixemes are, therefore, precisely the attributes that any colored entities of the two-dimensional geometry have, as well as the relations that they can assume among each other. In addition to the characteristic properties of geometric entities – in particular topological, metric, and directional relations between their parts (shape) and to other areas (position) – and the color markings in the narrower sense, which are more precisely determined by hue, brightness, and saturation, homogeneous color gradients or particular color variations – textures – can also be relevant as attributes of a higher order. In addition, interactions occur that result from the spatial arrangement of different colors in relation to each other, especially contrast effects. Combinations of pixemes, maximal pixeme Since the subdivision into​ »word«​ and​ »sentence«​ is irrelevant for morphological analysis of images, combinations of several pixemes can also be regarded as pixemes without further ado: The morphology of images thus consists of part-whole orders of pixemes, which span – mostly in several stages – between the image as a whole – as maximal pixeme – and the areas considered as minimal, each with only a single homogeneous marker occupancy. Therefore, pixemes of a higher order not only have a marker occupancy in the sense mentioned above. Instead, they have a quasi-pictorial substructure. For example, in Figure 1, the centrally located ribbon-like pixemes filled with brown and outlined in yellow form a columnar complex higher-order pixeme. Its basic geometric structure is not simply marked by color or texture values but precisely by the lower-order pixemes that constitute it. Although the pixemes of the “medium” order resemble an image carrier in their morphological structure, they cannot yet be used as an image carrier without further ado. That is mainly due to two related factors: a) In terms of Gestalt theory, pixemes always form figures: The background against which they are inevitably viewed as such does not belong to them. In the above example, the red areas enclosing the yellow-brown bands are not included. Although the pixeme segmentation can, in principle, be broken down into a variety of figure-ground pairs, the image carrier is only a figure in one respect, namely in front of the frame. Of all the pixemes involved, this only applies to the maximal pixeme and has a special effect there. Figure 3: Pi­xeme section as an image b) The framing of the maximal pixeme puts the latter in the context of use that makes this figure as a whole a sign mark in a sign action, i.e., an image carrier. Of course, in principle, it is quite possible to carry out this framing action with each of the lower-order pixemes, i.e., to regard them as separate image carriers (and thus as other images). However, with such an approach, the pragmatic and semantic references are not preserved: If one cuts out one of the yellow-bordered, brown-filled bands from the middle part of Figure 1 and mounts it alone on the background of a neutrally colored surface (or also freely floating in space), the result can certainly be understood as a picture with a somewhat unusually curved edge (i.e. a framing without an explicit frame) (Fig. 3). However, contexts of use and attributions of meaning to this image are at best very loosely related to those in Figure 1. Pixeme-forming operations Ultimately, the pixeme-​forming operations depend on the elementary pixeme attributes. Based on the psychophysiological mechanisms of perception, they usually take place unconsciously. Two opposing aspects are particularly significant here: contrast enhancement and Gestalt formation. The constitution of pixemes when looking at an image carrier is – as a variant of segmentation when seeing in general – strongly context-sensitive: Locally effectual contrast-enhancing components of the perception apparatus lead to boundaries between areas perceived as uniform. These operations become particularly noticeable when they lead to illusions, i.e., to additional pixemes (or more generally, perception segments), such as in the case of contrast illusion. On the other hand, the mechanisms of contrast enhancement support the merging of homogeneous areas through Gestalt formation in the sense of the Gestalt laws. Their unconscious activity determines the perceived part-whole hierarchies of the assembled pixemes. The interplay of drawing boundaries through contrast enhancement and integration according to the Gestalt laws ultimately leads to the constitution of an image-morphological (normal) structure between maximal pixemes and elementary areas, which, however, in everyday image perception is very strongly determined by semantic and pragmatic boundary conditions. It is precisely for this reason that design textbooks, for example, always emphasize that “the eye” in the design perspective must be trained in a creative way of seeing that ignores such influences and ultimately wants to achieve a purely intrinsic value approach to image morphology (e.g. [Klee 1956a]Klee, Paul (1956): Das bildnerische Denken: Schriften zur Form- und Gestaltungslehre. Schwabe, Basel, Stuttgart. Nicht angezeigt: bearb.: Jürg Spiller ). Applications Figure 4: Result of an automatic segmentation: texture-based pixeme formation. The pixemes found are color-marked on the right. The algorithmic imitation of pixeme-forming operations leads to the possibility of image morphological analyses in computer visualistics. It forms a central part of digital image processing: Corresponding segmentation methods can be programmed that operate on computerized encodings (i.e., notations) of image carriers and allow determining pixemes (often referred to as 'objects' in the relevant literature, cf. Fig. 4). Here, above all, the Gestalt laws of proximity, similarity, and good continuity are operationalized via the color and texture markers. In the long term, providing image science with a set of standard technical tools for morphological image analysis may be possible. That is particularly useful insofar as the pixeme composition of the image carrier, as mentioned above, has to be related to a theoretically assumed standard viewer. However, it must be remembered that the human perception of images, as described by Saint-​Martin, among others, includes not only possible individual deviations from the standard observer but also dynamic aspects that remain hidden by such a purely structural analysis. On a conceptual level, the synthesis of the pictorial morphosyntax from the basic geometric calculus and the conceptual field of the colored marker values ​​finally allows a long-held assumption to be refuted: namely that the concept of syntactic (non-)well-formedness could not be applied to images at all (cf. [Plümacher 1999a]Plümacher, Martina (1999): Wohlgeformtheitsbedingungen für Bilder?. In: Sachs-Hombach, K. & Rehkämper, K. (Hg.), Bildgrammatik. Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg. ). Since damage to the image carrier can disrupt the basic geometric structure of the maximal pixeme, the concept of a syntactically incorrect image makes perfect sense. See also: Color as an image-syntactic category Color perception Composition Digital image processing Framing, frame Gestalt Iconic difference Identity criteria for image carriers Image grammar Image perception Interaction and communication Morphology and syntax Notation Pragmatics, semantics, syntax Sign, sign carrier, sign system Space and geometry Syntactically incorrect images Texture Theories of image space Theories of visual perception

Notes Literature [Chomsky 1957a]Chomsky, Noam (1957): Syntactic Structures. Mouton, Den Haag.[Dölling 1999a]Dölling, Evelyn (1999): Kategorialstruktur ikonischer Sprachen und Syntax der visuellen Sprache. In: Sachs-Hombach, K. & Rehkämper, K (Hg.), Bildgrammatik: Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg.[Klee 1956a]Klee, Paul (1956): Das bildnerische Denken: Schriften zur Form- und Gestaltungslehre. Schwabe, Basel, Stuttgart.[Plümacher 1999a]Plümacher, Martina (1999): Wohlgeformtheitsbedingungen für Bilder?. In: Sachs-Hombach, K. & Rehkämper, K. (Hg.), Bildgrammatik. Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg.[Sachs-Hombach 1999a]Sachs-Hombach, Klaus (1999): Gibt es ein Bildalphabet?. In: Sachs-Hombach, K. & Rehkämper, K. (Hg.), Bildgrammatik. Interdisziplinäre Forschungen zur Syntax bildlicher Darstellungsformen. Scriptum, Magdeburg.[Saint-Martin 1990a]Saint-Martin, Fernande (1990): Semiotics of Visual Language. Indiana University Press, Bloomington, IN. Version 1: 2013 Responsible: Schirra, J.R.J. Editing: Sachs-Hombach, Klaus Contributions by: Joerg R.J. Schirra und Marcel Lemmes — (Note) Quote as: Schirra, Jörg R.J. (2013). Image Morphology. (Ausg. 1). In: Schirra, J.R.J.; Halawa, M. & Liebsch, D. (Hg.), Glossar der Bildphilosophie. (2012-2026). [bildglossar.org/index.php?title=Glossar-English:Image_Morphology].