Project Detail

Reconstructing George Washington

3D Face and Body Models of George Washington

Lead: Razdan, Anshuman
Collaborators: Cooper, Eugene;Farin, Gerald;Hansford, Dianne
RA: Tocheri, Matthew; Vannote, Scott;Hansen, Jeremy
Sponsor: Mount Vernon
Date: 05/25/04  - 02/28/06
Web site: http://prism.asu.edu/research/washington.php

Abstract

This project brought together forensic anthropologists, digital artists and computer scientists in the 3D digital reconstruction of George Washington at ages 19, 45 and 57, including detailed heads and bodies. Although many other scanning projects such as the Michelangelo project have successfully captured fine details via laser scanning, our project took it a step further, i.e. to predict what that individual (in the sculpture) might have looked like both in later and earlier years, specifically the process to account for reverse aging. Our base data was Washington's face mask at Morgan Library and Houdon's bust located at Mount Vernon, both done when Washington was 53. Additionally, we scanned the statue at the Capitol in Richmond, VA; various dentures, and other items. Other measurements came from the personal clothing and even portraits of the first President of the United States.

3D comparison of the Morgan life mask and the Houdon's bust scaled by 9% to accomodate shrinking of the terracota material during firing. Light green to light blue range of colors shows high simmilarity between the two.

The first few months of the project were devoted to scanning various artifacts in different parts of the country, followed by careful data processing and stitching to generate 3D images which were then used by the anthropologists to derive the face geometries at the appropriate ages.

The first step of the George Washington reconstruction effort involved acquiring 3D data that best represented the first president's physical characteristics. As direct access to Washington's physical remains was not possible, other evidence was carefully selected based on the potential 3D information that would be provided to the reconstruction effort. A plaster mask of the first president's face was the most important piece of evidence with respect to Washington's facial features and underlying bony structure. The mask is an actual mold of Washington's face, taken when he was 53 years old by the French sculptor, Jean Antoine Houdon. Currently, it is housed at the Morgan Library Museum in New York. Thirteen laser scans were taken from various angles to capture the 3D structure of the mask. Originally, Houdon used the plaster mask along with measurements he took of Washington to sculpt a bust as well as a life-sized statue. The bust is now part of the Mount Vernon estate while the statue is displayed in the state capital of Virginia in Richmond. The bust provided 3D data on Washington's head and neck while the statue provided 3D data on body proportions. Approximately 120 laser scans were taken of the bust and well over 500 scans were taken of the statue.

With 3D information about Washington's face, head, and body provided by the three aforementioned objects, evidence pertaining to his teeth and jaws was still required. A total of four sets of dentures (2 real, 2 replicas) were laser scanned to provide this information. The first set, housed at the Mount Vernon estate, consists of upper and lower jaw components. This set of dentures was composed largely of metal which complicated the scanning process because of the reflectivity of the surface. The remaining sets of dentures, housed at the National Museum of Dentistry in Baltimore, scanned more successfully. In particular, the resulting 3D model of a replica of the lower jaw denture from the Morgan Library Museum became vital to the reconstruction effort as it is a denture that Washington likely wore for a significant time of his later life. In general, well over 100 laser scans were taken on each set of dentures. In all cases, the research team traveled to the location of the object and laser scanning was completed on site under direct supervision of curatorial staff. The conditions for laser scanning were not always ideal and the team made adjustments as necessary to capture as much of the vital 3D information as was possible given the circumstances. All sets of dentures were scanned using the Cyberware Model 15 laser scanner whereas the life mask, bust, and statue were scanned using the LDI 100/200 laser scanners. Several additional items were also scanned (e.g., Corcoran mask, Mount Vernon medallion, etc.); however, these items did not contribute significantly to the reconstruction effort as their 'true' 3D relationship to Washington is suspect.

An example of how volume deformation tool was used. The jaw was deformed with the 3D scan of the face from Houdon's bust as a reference.

All the scanned data was pieced together using Raindrop-Geomagic software and the resulting data type was triangle meshes. The triangle mesh representation is the most common output of laser scanners and geometry software processing software. This posed another problem of editing and manipulating the geometry.

As would turn out to be the case with many of the "truths" about Washington or Washington representations, the first of these to fall was the oft-repeated story that Washington had not worn dentures when the mold for the life mask was cast in plaster. As was noted from photographs, and was later confirmed when digitally superimposing the 3D scan of the life mask with the 3D scan of the bust, there had to have been structure behind Washington's lips or the weight of the plaster would have notably depressed them into his oral cavity. Another result of this digital comparison was that the life mask and face of the bust were so similar metrically that it seemed eminently reasonable to use the bust in its entirety as a reasonable representation of the man's head in general.

The second sources of information that was needed in order to pursue the aging and especially the de-aging of Washington were his dentures. The only available complete set consisting of associated uppers and lowers are those in the collections of Mount Vernon. Differing from dentures typical of the 18th century in that lead rather than ivory or bone was the substrate in which teeth (in this case animal incisors in the upper and various human teeth in the lower plate) were fastened, they were also unusual in that the non-dental surfaces of both plates were essentially flat, rather than grooved to embrace toothless gums. Nevertheless, until scans of other dentures were acquired, the scan of the Mount Vernon dentures was used initially as a guide for digitally whittling to fit a scanned mandible of a French and Indian War soldier that approximated the width of what was determined from the digital bust and life mask Washington's mandible had been. Since the Mount Vernon denture possessed a hole to receive Washington's last tooth - a left lower left second - the appropriate tooth was left in the digitally altered mandible and adjusted to height of the teeth in the lead plate. After an exact replica of the surviving lower plate of the dentures Joseph Greenwood made in 1789 for Washington was scanned, the digitally manipulated mandible was further adjusted to fit the groove along this plate's underside. When the senior anthropologist was satisfied with the alveolar curve of the digital mandible, the 3D data was placed in the bust in order to adjust the angle of the chin, which, in virtually all representations of Washington as an older individual, is obliquely slanted down from left to right. Having thus reconstructed a reasonable facsimile of Washington's mandible at the age of 57 (which he would have been in 1789), the fit was fine tuned by adjusting known average skin depths for older individuals at crucial point along the lower jaw. Reconstructing the upper jaw was less straightforward. When initially inquired about the 1795 Greenwood dentures, Scott Swank, curator at the National Museum of Dentistry, noted that although the original had been stolen from the Smithsonian in the 1980s and only the lower recovered, there was a replica. Upon seeing the replica, it was immediately obvious to the anthropologist that it did not capture the shape and size of the original, much less its details. Fortunately, Swank was able to provide a photograph that portrayed both the original and the replica. As such, it was realized, once scanned, we could use the replica upper and lower as a scale by which to calibrate the size of the original. Subsequently, the photograph was scanned and the images of the original and replica upper superimposed in order to achieve a curvature that could be used to remodel the scanned upper jaw of the soldier whose mandible had been used. When the toothless digital maxilla seemed satisfactorily modified, it was placed into the digital bust and skin depth adjusted accordingly. This then provided the bony infrastructure for Washington's face at the age of 57.

Final results of facial reconstruction of George Washington at ages 19, 45 and 57.

In order to achieve the face of the 45 year old, we looked at Charles Wilson Peale's portraits of Washington at the ages of 40 and 47, which visually represented an individual who lower face was longer from nose to chin, and whose chin was rounded symmetrical. Given Washington's self- proclaimed proclivity for cracking walnut shells with his teeth, and the common loss of posterior teeth prior to front teeth, the longer lower face could be explained by the retention of anterior teeth. In order to test the visual observation, prints of both portraits were scanned and relevant facial features digitized. The two scans were then compared and, indeed, they were similar in length of the lower face and symmetry of the chin. The two-dimensional scans of the portraits were superimposed over the bust, which was oriented in the position of the pose. Upon aligning facial features, it was indeed the case the 40 and 47 year old faces were longer from nose to chin, and the face of the bust was adjusted accordingly, including rounding up the inferior margin of the chin. The digital mandible and maxilla were then placed into this remodeled face and adjusted to it, using average skin depth measurements. The angle of the mandible was left essentially unaltered from its amorphous shape in the bust and portraits in accordance with the hypothesis of posterior tooth loss and concomitant bone loss. The flesh around the neck was also tightened up a bit.

Having thus lengthened the lower face, the reconstruction of the lower jaw of the 19 year old required manipulating the back of the mandible to assume the right-angled look a pre-tooth loss jaw of a young individual would have had. In turn, the inferior margin of the body of the mandible had to become more horizontal and it and the angle made more prominent. Since cartilage grows throughout one's lifetime the tip of the nose and its alar cartilages as well as the earlobe were shrunk (exposed below the hair in the life mask). Hypothesizing that Washington might have continued to acquire the somewhat more pronounced supraorbital and glabellar regions seen in the older adults, the anthropologist required that these regions be flattened. A younger appearance was also achieved by removing wrinkles and adding a bit to the cheeks, which would have lost fat with age. The skin of the neck was tightened further. It was also decided to fill in the pock mark under the left zygomatic arch that is seen in older representations of Washington and which could have resulted from the bout of smallpox he suffered when he traveled to Barbados with his brother later that year.