Recent research applications of biostereometrics embrace more powerful faster and more cost/effective technologies for sensing recording analyzing and displaying spatial and spatio-temporal information about biological form and function. As well as providing more systematic basic biological data these developments have led to more efficient precise and effective surgical procedures and computer-aided sculpting of body parts. Recent educational applications include the production of interactive 3D atlases artificial reality simulations and knowledge bases with statistically-based information about individual organs and tissues.

This project is based on an entirely new concept for teaching the structure and function of the human body a concept which combines traditional approaches gained from centuries of study of human anatomy with the most recent sophisticated 3-dimensional computer graphics display systems and laser disc technology. The end-point of the project is a high resolution interactive 3-D atlas of human/animal anatomy stored on a laser video disc and displayed on graphics workstations--an " electronic Gray''s Anatomy" . These displays will be used to teach the structure of the body and to give students and instructors an understanding of their own body in health and disease. To evaluate the software developed undergraduate students from the anatomy courses at CSU wil be allowed to work with the computer-generated images from the earliest stages of development. Feedback from these students will be incorporated into the software development. Furthermore once a relatively complete series of images has been generated groups of students will be selected at random to study anatomy with this new methodology and will be compared with control groups who utilize more traditional techniques. METHODOLOGY This is a complex project that requires many individual facets to be developed simultaneously (figure 1). We have established an important collaboration with the Uniformed Services University in Bethesda that will allow us to utilize a large cryotome with photographic systems and the expertise to operate it already available there. Indeed most of the elaborate apparatus such as graphics workstations needed for the project is currently available either at CSU or through collaborative arrangements with other institutions.

The advantages of using three-dimensional (3D) data in the description and analysis of biological forms are obvious: these data provide realistic geometrically integrated models of the forms under study and can be rotated translated and dissected electronically for viewing. 3D coordinate data can be collected from several sources including computed tomographic images stereo photographs specially designed microscopes and digitizers. But once collected how can these data be analyzed to address biologically relevant research questions? This paper demonstrates the capabilities of two analytical techniques finite-element scaling analysis and Euclidean distances matrix analysis in the comparison of 3D biological forms. Examples include studies of growth of the craniofacial complex and analyses of differences in form between members of biologically defined groups (e. g. species sexes diagnostic categories).

Extensive developments have taken place in three-dimensional Magnetic Resonance Imaging technology. These innovations make it possible to perform complex mathematical analyses in order to evaluate two- and three-dimensional data sets. Our purpose is to describe some of the emerging array of MRI acquisition processing and display techniques and to demonstrate their applications to questions about functional and evolutionary anatomy of musculoskeletal components of the head.

Anthropologists and paleontologists have traditionally relied upon instruments such as calipers to measure bones and teeth thus effectively reducing complex threedimensional surfaces to two dimensional characterizations of surfaces. The Reflex Microscope is a noncontacting threedimensional measuring instrument that allows easy computation of measurements such as volumes and complex surface areas through its accompanying computer software. In the present study high resolution dental impressions were taken every 68 weeks from a laboratory colony of 15 vervet monkeys. Epoxy casts of the teeth were analyzed using the Reflex Microscope. Results indicate that monkeys raised on different diets show different wearrelated changes in tooth shape.

A low cost real time method of measuring the movement of the ribcage and abdomen during anaesthesia is described. The equipment comprises a scanning light stripe system video pre-processing electronics and a personal computer. Selected chest surface contours are measured at the rate of 2Oms per contour. Linear interpolation is used to provide contour area estimates between contour sample periods to allow chest volume to be calculated at 2Oms intervals. Results using test objects show that the equipment is able to measure volume to an accuracy of beuer than 1 with reproducibility to within 0. 3

Geometric measurements of a surface can be encoded in real-time as a set of fringe images using moire projection techniques. However obtaining numerical values from the encoded surface measurements has not been straightforward. To solve this problem we have developed a phase-shift moire camera that captures in real-time a sufficient number (three) of phase-shifted moire patterns to allow the decoding of the moire patterns to be automated. The RIPS Camera uses three simultaneously produced phase-shifted moire fringe images of a scene (above left) to produce data (above right) that can be reduce to range data.

Since the development of the ophthalmoscope ophthalmologists evaluated subjectively the retinal ocular fundus for abnormalities and changes that occur in glaucoma and found that thinning or loss of retinal nerve fiber layer (NFL) thickness may preceed visual field loss caused by glaucoma. As a quantitative and objective technique to detect glaucoma at an early stage as possible we have applied a photogrammetric technique to measure NFL thickness along the optic disc margin and peripapillary region using stereophotographs taken with the Donaldson fundus stereocamera.

The MetaMap process extends the concept of direct manipulation human-computer interfaces to new limits. Its specific capabilities include the correlation of discrete image elements to relevant text information and the correlation of these image features to other images as well as to program control mechanisms. The correlation is accomplished through reprogramming of both the color map and the image so that discrete image elements comprise unique sets of color indices. This process allows the correlation to be accomplished with very efficient data storage and program execution times. Image databases adapted to this process become object-oriented as a result. Very sophisticated interrelationships can be set up between images text and program control mechanisms using this process. An application of this interfacing process to the design of an interactive atlas of medical histology as well as other possible applications are described. The MetaMap process is protected by U. S. patent #4

Determination of a target position requires the obtaining of x-ray source position and image plane orientation in 3D space. With those parameters target position is determined from the intersection of two lines joining the x-ray source and the target image point in space. The error of target localization is represented with an analytically derived covariance matrix. The covariance matrix from this analysis is verified to be in good agreement with simulation.

We discuss automatic acquisition of movement information from two types of input data using a knowledgebased scheme. When the input data is an image of 3D surface points mensurated from a single perspective the process of recognizing and locating a body position in each image frame includes (1) isolating the body from the motion scene (2) segmenting it into the individual body parts and (3) identifying these parts with the labeled body parts of a stored model. When the input data is a limited number of 3D points indicating key body parts in activities only step (3) is required for movement recognition: we only need to recognize light emitting devices (LED''s) attached to the reference points without going through every detail of body surface points. The constraints in the system are explored and propagated to extract structural information from the input data compatible with the model structure.

In this paper moire topography is applied to study the femoral articular facies of the knee of Primates. For compari son with each other of different families of Primates we suggest the comparative targets a y and the grade G of the moire contour fringes on two condyles of knee of Primates and comparative study of the articulation of knee between the Macaca assamensis M cellaud Presbytis phayrei Rhinopithecus roxellanae Hylobates concolor leucogenys Nycticebus concany Gorilla gorilla Anthropopithecus troglodytes Sirnia satyrus and human being are given. The results may be useful reference in the study of Biomechanics Zoology and Anthropology.

The segment mass mass proportions and moments of inertia of a sample of twelve females and seven males with mean ages of 67. 4 and 69. 5 years were estimated using textbook proportions based on cadaver studies. These were then compared with the parameters calculated using a mathematical model the zone method. The methodology of the model was fully evaluated for accuracy and precision and judged to be adequate. The results of the comparisons show that for some segments female parameters are quite different from male parameters and inadequately predicted by the cadaver proportions. The largest discrepancies were for the thigh and the trunk. The cadaver predictions were generally less than satisfactory although the common variance for some segments was moderately high. The use ofnon-linear regression and segment anthropometry was illustrated for the thigh moments of inertia and appears to be appropriate. However the predictions from cadaver data need to be examined fully. These results are dependent on the changes in mass and density distribution which occur with aging and the changes which occur with cadaver samples prior to and following death.

Growth data on the developing human face has been collected by scientists in a number ofdifferent disciplines for nearly one hundred years. Most of the work in this area has been the result of attempts to understand and quantify the growth process or in attempting to establish population norms or baseline standards for use in syndromology. Anthropologists geneticists ophthalmologists orthodontists otorhinolaryngologists pediatric syndromologists plastic surgeons radiologists have all contributed to the literature on facial development1. While this literature for facial assessment is very complete but its use and value in prediction has not been exploited to-date. The preliminary work presented in this paper uses the available literature to predict growth and development ofthe nose. These predictions were used in fabricating a series ofnasal prostheses that will be used to establish facial cosmesis during the growth years. This case involves a six year old male who had undergone a total rhinectomy and septectomy required a nose prosthesis to cover a severe facial defect and to effect cosmesis. Based on laser scan data from casts ofthe patient''s pre-surgical nose a selected series ofanthropometric landmarks were identified and scaled. Growth algorithms provided new positions for control points that simulated growth of the nose over time. The resulting series of geometric models were initially fabricated using stereolithography and later modeled in a medical grade silastic. Growth predictions used to simulate the morphological remodelling of facia

The reconstruction of facial features on the human skull for purposes of identification has, in the past, utilized either two-dimensional drafting or three-dimensional sculpting techniques. Two and three dimensional computer aided routines are being developed to minimize errors introduced by limits of artistic ability or inconsistencies in the application of techniques. These routines allow generalized facial features to be manipulated via control points to conform to the size and shape of a specific skull. Subtle alterations of the surface form, texture and color based on age, sex and race enhance the individuality of the generated facial form.

Two-degrees offreedom laser scanning devices offer the advantage ofspeed in data acquisition over three-degrees offreedom devices. Data acquisition on patients especially children requires speed in data gathering instantaneous verification ofdata integrity and more than two degrees offreedom to adequately describe morphology. A technique to generate a pseudo-third degree of freedom scan will be presented. An additional benefit of the technique is the ability to compare data acquired at different times.

Good quality measurement of the exposed surface of a complexshaped object requires ix mechanical degreesoffreedom (DOF5) between the object and a ranging system. Unlike many conventional digitizing systems which provide only four of the six DOF5 the DIGIBOT system provides five DOF5. Each additional DOF above four requires increasingly sophisticated mechanics control and data processing. With its five DOF5 adaptive/closed-loop scan control and enhanced data processing capabilities the DIGIBOT system has been used to measure relatively complex anatomical subjects. 1

In 1979 we first reported progress on a unique 3-D quantitative mapping structured light method to rapidly and remotely interrogate an unknown surface with an array of laser beams and to observe and measure the 3-D surfaces off-axis with one or more videocameras. -''2 Significant progress has occurred recently to make this method practical and cost-effective. The system consists of: 1) at least one passive sensor typically a 512 X 512 square pixel lowlight-level CID image intensified progressive scan videocamera operating at 30 Hz 2) an off-axis active structured laser projector using a 532nm doubled YAG solid state polarized laser beam formers and a prismatic optic to simultaneously generate a 128 X 128 square laser beamlet array 3) a spatially programmable light modulator to block or transmit selected sets of 128 rows (or 128 columns) of laser beams (corresponding to the generated array) at millisecond frame rates 4) optics to project the structured light sets onto the object 5) optics at the sensor end 6) positioning equipment to manipulate the object as required 7) suitable image processors array processors and computers for signal and image processing calculation of tables of X numerical data interactive display and archiving 8)interfaces to other CAD/CAM systems for the replication of surfaces or the reconstruction of multiple surfaces into geometric forms. .

A 3-D surface is mensurated by active triangulation: an electro-optically controlled laser camera projects an array of beams onto the surface and an optical camera detects the reflected images from a different perspective. By controlling the projection of the rays in a binary pattern mode it is possible to correlate automatically the surface positions illuminated by the active camera and observed by the passive camera and thus achieve the rapid triangulation of multiple surface points in three dimensions. The present implementation technique is discussed in comparison with other data acquisition methods. After an entire body surface i mensurated by multiple perspectives around the body a volumetric representation is generated. Some examples are displayed using computer graphics.

A method based on differential geometry is presented for quantitatively measuring facial (or body) surfaces and the amount of change produced in them by facial surgery. Facial shape is described mathematically in terms of 8 fundamental surface types. Changes produced in the shape of the face are described in terms of changes in these constituent surfaces. The method is illustrated using examples of faces changed by reconstructive surgery and its relationship to clinical observation explained.

An integrated system for the planning of maxillo-facial and cranial surgery is described. The system is based on the use of computer graphics to simulate surgical operations on the bone and to predict and present the outcome of these on the soft tissue and appearance of the face. The input data consists of a suitable set of contiguous X-ray CT slices. As well as the core system which provides the simulation and modelling an extensive range of support facilities is provided for measuring external and internal anatomy profile plotting and quantitative postsurgical assessment. A system for long term follow up of facial shape change is provide by an optical surface scanner. Facilities are also provided for the downloading of data to a numerically controlled milling machine for the production of customised prostheses.

The cosmetic result of an ear restored surgically or via prosthetics is dependent on the surgeon''s ability to carve a precise cartilage armature at the time of surgery or the prosthetist''s ability to sculpt in wax an exact duplicate of the patient''s " missing" ear. Introducing CAD/CAM technology into the process benefits the esthetic outcome of these procedures. By utilizing serial section information derived from CAT MRI or moulage techniques a mirrorimage of the patient''s " donor" ear is generated. The resulting earform data is then used for the design of a cartilage armature produced by multi-axis milling or to produce by stereolithography a model which serves as the basis for a prosthesis.

A generalized " anatomically standard" eyeball was geometrically modeled on a Computer-Aided Design and Drafting (CADD) workstation based on published conceptual norms of dimension radii of curvature alignments etc. An orbital environment for this model was concurrently prepared by serial section reconstruction of a cadaver specimen. Issues addressed include orienting unregistered sections the utility of interactive CADDsupported visual logic in interpretive decision making and the value of geometric models.

A simplicial algorithm using a door-in/door-out technique is developed for triangulating between parallel contours which bifurcate. Our immediate application is in biomedical imaging using surface modeling methods.

The biological mechanism for the formation of retinal vessel patterns in the developing human eye is unknown even though it is a question of importance. The current hypothesis is based on the existence of a variable oxygen gradient across the developing photoreceptors which stimulates the release of angiogenic factors which diffuse in the plane of the retina and result in the growth of retinal vessels. This implies that the rate-limiting step in the formation of the vessel pattern is a diffusion process. In order to analyze this hypothesis the fractal dimension of the retinal blood vessel patterns was determined. Several methods were used to calculate the fractal dimension. Red-free fundus images of normal humans were traced and two methods of analysis were used: the massradius relation and the scaling relation of the two-point densitydensity correlation. In addition the vessel patterns were digitized using a digital image processing system and the number of pixels corresponding to the retinal vessels were determined within circles of various diameters. The slope of the log of the size of the circle versus the log of the number of voxels yielded the fractal dimension. All three methods applied to retinal patterns derived from 15 different normal humans (age 14-65) showed that the human retinal blood vessels have a self-similar structure with a fractal dimension of about 1 . 7. This is the same fractal dimension found for

Atechnique is proposed which uses fractal analysis for the non- traumatic and non-invasive quantification of trabecular bone density in the mandible using standard dental radiographs. Binary images of trabecular bone patterns are derived from digitized radiographic images. Fractal analysis is then used to calculate the Hausdorif dimension (D) of the binary image patterns. Variations in D calculated with this method can be correlated with known cases of systemic osteoporosis to establish normal and abnormal ranges for the value of D.

Failure of the interface between implant and bone and consequent sinkage or loosening is recognised as the major cause of failure of artificial joints. It is commonly suspected that the ultimate lifetime of the joint may be indicated by early patterns of such movement but these are at or beyond the limit of conventional radiography to resolve. Several groups have applied photogrammetric methods to the problem. But while the absence of optics allows straightforward analytical solutions practical difficulties abound such as degradation due to X-ray scattering from the large object support of the infirm patient landmark identification on the bone and implanted components and speed and precision of film measurement. This paper introduces these difficulties and presents the various approaches to their solution currently being adopted for clinical trials at the Nuffield Orthopaedic Centre.