THE MORPHOLOGY PROJECT ART-SCIENCE EXPLORATIONS OF BIOLOGICAL SHAPE ANALYSIS AND EVOLUTION

By ROB O’NEILL

INTRODUCTION

Morphology, within the field of biology, refers to the outward appearance of an organism or taxon and its component parts. As a means of quantifying morphology, morphometrics is the collection of methods used to collect measurements from shapes and perform statistical analysis on the variation. This paper introduces the concepts and history behind morphology and morphometrics as a backdrop to the subsequently described art-science explorations undertaken by the author in the Morphology Project. In particular, “dataFace”, the most recent installment of the Morphology Project is detailed as an example of how a large scientific dataset can be harnessed using custom and existing tools as a means of artistic exploration and scientific inquiry. Morphometrics is an untapped area for the arts. By utilizing these ideas and their accompanying methodologies, artists have the ability to analyze shape mathematically and to wield large biological shape databases as a medium. This interplay has the potential to enhance both the arts and sciences.

BRIEF HISTORY

In the second half of the nineteenth century, the use of numbers and mathematics swept through the human and biological sciences. For the first time, descriptive subjects could add a layer of quantifiable data in the same manner that physicists could rely on mathematics to support their claims (Gould, 1996). It was not long before these numbers, particularly those derived from craniometrics, were being used to justify the social beliefs of the time. Craniometrics is the measurement of the human skull. During the late nineteenth century, the most common measurement was that of brain size most often recorded from the volume of the neurocranium or braincase. The calculated cubic volume of the brain was a single variable that could be routinely recorded and individual values easily sorted into a list. As the seat of intelligence and thought, the brain was considered a likely distinguishing characteristic that could be used to separate people into divisions of class, intellect, race, gender, or whatever else was fashionably in need of a hierarchical ranking of individuals and groups. Factors that influence brain size, such as the individual’s height, were rarely used to equalize the data. The sensation that craniometrics created prompted living scientists at the time to wager between each other as to the ultimate size of their posthumous brain volume and the recording of brain size of notable intellectuals became a common practice and recorded item. This work is often associated with French physician, anatomist, and anthropologist Pierre Paul Broca. When measuring the brains of human skulls was not enough data, the calculations required to collect brain volumes from living people was developed and incorporated into the study of anthropometry. The measurements were considered to be indicative of inequalities derived through nature or divine plan, but also predictive of an individual’s tendencies.

At the turn of the nineteenth century, Cesare Lombroso popularized the notion of the “born criminal”, claiming that criminals have particular physiognomic attributes or deformities. Physiognomy attempts to estimate character and personality traits from physical features of the face or body. To Lombroso, the violent criminal had devolved, and therefore criminals were societal or evolutionary regressions, retaining primitive traits. If criminality was inherited, then the “born criminal” could be distinguished by physical atavistic stigmata, such as large jaws, forward projection of jaw, low sloping forehead, high cheekbones, flattened or upturned nose, handle-shaped ears, hawk-like noses or fleshy lips, hard shifty eyes, scanty beard or baldness, insensitivity to pain, and/or long arms (Gould, 1996).

During the 1920s and 1930s Franz Boas and other anthropologists began to use anthropometric approaches to discredit the concept of fixed biological race. While at the same time, Nazi Germany famously relied on anthropometric measurements to distinguish Aryans from Jews during the Holocaust. When these atrocities were discovered one of the effects was the bright light cast on the insignificance of anthropometric techniques and the eugenic ideals they were supporting. Anthropometry generally went into decline starting in the 1940s because of these events.

Over time, the complexity of statistical operation and the techniques for capturing measurements expanded, but more importantly the intentions of the research evolved. Significant data was captured to understand the nuances of species variation and for understanding the population as a whole as opposed to the ranking or isolation of an individual or group. Craniometrics and physiognomy were replaced by morphometrics, biometry, and biomathematics. Today, morphometrics is widely practiced, as a means of collecting and analyzing data on biological shapes. Morphometrics is based on shape, which is invariant to translation, rotation, or scale. Form, on the other hand, can be considered shape plus scale. From insect wings, to the profile of eye sockets, to the overall shape of long bones, the methods are expanding and the sensitivity to the signal of a species is becoming more attuned. Passing are the calipers and custom measuring devices, replaced by laser scans, software, and other means of digitizing, storing, sharing, and analyzing increasingly large datasets. Morphometrics is helping researchers collect and share complex data and to illuminate and give form to the connections between species illustrated by the mounting wealth of molecular data.

THE MORPHOLOGY PROJECT

The Morphology Project is a series of experiments designed to merge the study of comparative biology, from the perspective of morphology, with the methods and tools of digital animation. The outputs range from techniques for scientific analysis to art projects that educate and provide another level of meaning to the large datasets being processed. Projects have included: “The Theory of Transformations”, a modern take on D’Arcy Wentworth Thompson’s grid-based species warping theories (Thompson, 1917); “dataProjections”, an artistic rumination of contemporary morphometric data and the gesture of scientific data collection; and “dataFace”, a modern look at craniometric human datasets. All of these projects have been created with data collected by practicing scientists, usually anthropologists, for their own purposes, then re-imagined artistically by the author. The emphasis is on artistic visualization of scientific data.

In “The Theory of Transformations” (2006), laser scans of a modern human and chimpanzee crania are three-dimensionally warped to resemble each other via procedural operations on a deformation grid informed by D’Arcy Wentworth Thompson’s 1917 work (Thompson, 1917). Thompson presented a human cranium and the associated Cartesian warped grids that would resolve a chimpanzee and baboon skull. “The Theory of Transformations” takes his grids and matches them identically to transform a human skull, with its perfect but biased square grid, to the sweeping curved grid of a chimpanzee. The resulting form is similar to that of a normal chimp, but not, as would be expected, the perfect specimen resolved by Thompson. When a real chimp scan is placed into the grid and unfolded to reform the human square, the result, while not human, is reminiscent of an immature chimpanzee which are always likened to humans. These transformations are presented as a looping movie on a small LCD monitor that is embedded into an antique wooden incubator in the tradition of a “Cabinets of Curiosities” (also know as a Wunderkammer) that also contains the rapid prototyped 3d print of the chimp to human model. By presenting this as an animation, Thompson’s theories are taken to their mathematical extreme and we see the shapes in between the end goals. In three-dimensions, we are able to do this deformation more accurately, see the structures from infinite perspectives, and watch the process over time. Looking at Thompson’s work and reinventing it, we gain a deeper understanding of the comparative biological shapes in our world. While Thompson’s theories have wavering scientific relevance, the forms produced between the start and end points are where they become most interesting. These forms are the ones Thompson never saw. Beginning with the realistic starting points of laser-scanned chimpanzee and human crania, we apply the others transformation grid and travel through hypothetical forms to reach something created through sculpture and code that has the semblance of the others form, appearing both familiar and unreal.

In “dataProjections”, 3D data collected by a biological anthropologist is traced, recreating the path of the scientists hand and creating a model structure reminiscent of the gorilla crania mapped. “dataProjections” is a print series depicting 3D structures generated procedurally from morphometric data collected in the process of research by a biological anthropologist on a series of gorilla crania. Passing are the days when a scientist would measure skeletal elements with calipers. In its place are digital tools that capture not only the shape of the specimen in three-dimensions but also the gestural path of the researcher’s process. Modern morphometric analysis is carried out via anatomical landmark location and recording by the biologist using a stylus-like apparatus that notes the three-dimensional Cartesian coordinate of a point on the specimen to a computer. This process is akin to the scientist invisibly drawing the specimen and is indicative of the seldomly represented gestural process of science. These three-dimensional points are stored in a spreadsheet or database and statistical operations are performed on them to analyze the shape. A suite of small custom scripts was written to import, analyze, and visualize this data in Autodesk Maya. Animation software such as this is not commonly used in the biological sciences, but through the process of creating this series of work a tool set for geometric morphometric analysis was developed. For this piece, a procedural curve is drawn through the imported point cloud, following the path of the scientist’s hand in the collection process. Visual reconstructions from this data are typically simple, but these images attempt to reveal the beauty inherent in the landmarks and the specimens they represent. The structures generated were then either mirrored to create a semblance of the original object

(Figure 1) or placed to create a graphic shadow representing the projection of thought and analysis based on the data

(Figure 2). The graffiti-style result of the shadow is a representation of the scientist’s hand through the eyes of the artist.

All of these projects are carried out in animation systems such as Autodesk Maya and augmented by custom plug-ins that allow atypical data to be read, analyzed, and visualized using the power of Hollywood-style animation and computer graphics tools (O’Neill, 2007). As a detailed walk through of methodology, the “dataFace” project is described below. This project is the culmination of the techniques developed in the preceding projects and deals with the largest dataset of the projects developed thus far.

DATAFACE: AN EXAMPLE METHODOLOGY

The latest project, “dataFace”

(Figure 3), is part data analysis and mapping and part artistic experiment. The project utilizes the database of craniometric measurements collected by Harvard anthropologist W.W. Howells (1908-2005) and transforms them into a three-dimensional point cloud. Howells made 82 measurements to record distances between landmarks and angles defined by anatomy from the crania of 2524 individuals collected from around the world and housed at a variety of museums. This cloud is converted into a deformation surface and is used to recreate a semblance of the person whose skull was measured. Bringing life back to this data is a unique process and one that opens other artistic doors, namely applying evolutionary algorithms to the population to evolve this human sample thousands of generations. By applying parameters that attempt to recreate extreme environmental pressures, a virtual petri dish of human facial evolution is created. Notions such as mutation, adaptation, and emergence in a virtual evolutionary system are sought after. In addition, physical simulation is applied to the deformation hull to enact forces on the resulting populations over time. As previously noted. craniometric data such as this has historically been misappropriated and used for the purposes of racial classification. This project mixes the same data as a sample of the global human population and attempts to artistically evolve it.

The project started as Howells’ data stored as a series of text files housed and maintained by the Anthropology Department of the University of Tennessee, Knoxville (Howells, 1996). Individuals collected by Howells at the American Museum of Natural History were identified and laser-scanned to create three-dimensional models. One cranium was selected as the base model. All of the landmarks were created as points on a deformable mesh hull in Autodesk Maya and relationships were created between these landmarks and the measurements in Howells’ data. These relationships acted as input to custom software running inside Maya that positioned the landmarks appropriately. Distances and angles provide only relative information and no actual positional data. Therefore, an elaborate rig was created which successively updated the position of rigid landmark-pairs based on the position of other landmarks derived from subsequent measurements. Along the way, the landmarks drove the position of the points defining the deformation hull which in turn warped the laser-scanned crania. The resulting cranium is compared to the other laser-scanned crania to check for accuracy. A three-dimensional point cloud of the landmarks, similar to those collected by biologists today, was created and used for comparison. Once working, this system can reconstruct any of the individuals that Howells’ recorded. This collection of three-dimensional surfaces acts as digital population with parametric controls. If each of these controls is abstractly thought of as a unit of inheritance, the population as a whole can be programmed to evolve over a long series of generations. Weighted factors or physical trends can be inserted into the system to drive the shapes in different directions. While environmental forces that modify morphology are hard to identify at the level of the crania, and may not exist, certain pseudo-forces were applied as mutations to the procedural system. Rules can also be applied to the system so that constraints and factors related to external conditions can effect the population and the change over time. In developing this system there was a sense of peering into the diversity of the human species. While watching thousands of crania be processed, notions of mortality and romanticism often associated with the human cranium are replaced with the desire to ascribe faces and to reinstate the individual amongst the numbers. Research and work is going into utilizing a procedural forensic model to build faces on these derived crania but this process is fraught with pitfalls and is being approached carefully and slowly. The hope is that through careful creation the system could also be used as a database to generate variations for synthetic actors as one application in a realm of many possibilities.

This system provides the artist and scientist with a parametric database centered on the human crania. Data driven art often deals with the information around us, be it weather, financial, or historical. The numbers represent systems that speak to the world we live in. “dataFace” deals with the data that is in fact us. As it is a completely procedural system built each time it is executed, the possibility of extension opens the door for the introduction of new methodologies and datasets.

FUTURE

The future of the Morphology Project lies in real time applications. The next phase of work is to undertake live analysis on faces both of users and figurative statues as opposed to static or historic datasets. The “reFace” project analyzes figurative sculptures and deforms an animatable 3d model to fit its proportions. This animated face is then projected back onto the sculpture to create an impromptu performance and potentially an unannounced spectacle. It can be hypothesized that this project could be extended to capture the faces of viewers using facial recognition software then mapped to data existing in the system. This notion of real-time morphometrics could serve as yet a new database to be analyzed, reflected upon, and presented in the arts context. Now the artist is collecting the data relevant to the sciences, bringing this project full circle and continuing the conversation between art and science.

The project can be followed at: http://www.morphometric.com/morphology

References:

Howells, W. W. (1973) Cranial Variation in Man: A Study by Multivariate Analysis of Patterns of Difference Among Recent Human Populations. Papers of the Peabody Museum of Archeology and Ethnology. Volume 67. Cambridge, MA: Harvard University.

Howells, W. W. (1989) Skull Shapes and the Map: Craniometric Analyses in the Dispersion of Modern Homo. Papers of the Peabody Museum of Archeology and Ethnology. Volume 79. Cambridge, MA: Harvard University.

Howells, W.W. (1996) Howells’ Craniometric Data on the Internet. American Journal of Physical Anthropology, Volume 101:441-442. http://konig.la.utk.edu/howells.htm

Lombroso, C. (1876) L’Uomo Delinquente. Milan: Hoepli.

O’Neill, R. (2007) Emerging Congruence Between Animation and Anatomy. Leonardo, 40(2), 169-173.

Thompson, D. W. (1917) On Growth and Form: The Complete Revised Edition (1992). New York, NY: Dover.