![]()
Title, Supervisors, Qualification aimed for, Computer Animation background, Animating the Human Form, Aims of the Project, Expected benefits to Industry, Project Plan, Required Resources, References, Menu Bar.
Department of Mechanical Engineering and Member of the Computer Graphics Research Group, Department of Computer Science, University of Sheffield.
Dr Steve Maddock Lecturer in Dept of Computer Science.
Dr Ian Badcoe, head of R&D at Gremlin Interactive.
G.N.Rutty, Senior Lecturer in Dept of Forensic Pathology.
Additional Industrial support from Gremlin Interactive Ltd.
Project Proposal for the degree of Master of Science by Research MSc (Res) in Materials, Structures and Systems Engineering. In the academic year 1999 to 2000. Submitted under the guidelines of the Department of Mechanical Engineering
![]()
Title, Supervisors, Qualification aimed for, Computer Animation background, Animating the Human Form, Aims of the Project, Expected benefits to Industry, Project Plan, Required Resources, References, Menu Bar.
Computer animation has seen a rapid improvement since 1963 and Ivan Sutherland¡¯s creation of Sketchpad [MYER98]. Today computer animation has supplanted most of the more traditional animation techniques; in fact traditional animation terminology is still used in the modern world of computer animation. The de facto standard animation techniques of 'keyframing' and 'inbetweeners' were first introduced by Disney in the early days of cell based animation, [THOMF81]. This technique worked via the senior animators creating key frames while in between frames were drawn by junior animators. This controlled technique is still employed but with a computer application undertaking the time consuming task of creating in between frames, this mathematical calculation is called 'interpolation'.
The problem with automatic interpolation is that with complex entities like the human body this technique is not intuitive compared to its human counterpart. The results of such interpolations are rarely realistic with limbs lurching between keyframes with no inbuilt regard for human anatomy or biomechanics. If an approximate model of the human body and its movement could be incorporated into the process of interpolation then this should improve the ability to model the human form.
![]()
Title, Supervisors, Qualification aimed for, Computer Animation background, Animating the Human Form, Aims of the Project, Expected benefits to Industry, Project Plan, Required Resources, References, Menu Bar.
![]()
¾Æ´ãÀÇ ÈïÁ¤ |
The human body is a highly complex system. The main component that concerns animation is the musculo-skeletal system, this is made up of bones, ligaments, tendons and various forms of muscle, which collectively allows the human form to move. The human body contains around 200 bones, this number varies because some bones fuse at different periods of time and the number of rib bones differ between the sexes. Ligaments connect bones to bones, and tendons connect bones to muscles. Smooth and slippery cartilage covers the ends of the bones to reduce, but not eliminate, friction during the movement. The strength and elasticity of ligaments and tendons differ from person to person and the effectiveness of cartilage deteriorates through, previous injury, age and medical complaints.
Modelling biological complexity on such a scale requires approximations and simplifications to be assumed, as it is currently implausible to create an extremely accurate human model. To approximate even the most basic component of human movement requires the generation of extremely complex and costly algorithms using extremely complex models of human anatomy. Models such as these are used in various fields such as sports science, biomechanics and medical simulations despite their shortcomings. Many papers describe research into the areas of complex human modelling but are limited to simple movement or specific areas of the body, [BRO98, HEE95, HOD95, HOD96, HOD98, MAU96, ZOR99].
In order to realistically animate the whole human form with a degree of realism alternative methods needed to be employed. This is once more developed from traditional animation methods, this time from the 'rotoscoping' technique. Rotoscoping was invented in 1915 by Max Fleischer [MALT80] to animate 'Koko the Clown' in the successful cartoon series 'Out of the Inkwell'. This technique works by tracing complex animation details from cinema, photographic or video recordings of motion in the real world. This technique was, and still is largely used to act as a simple guideline for two-dimensional cell animation. This was because the strict use of rotoscoping clashes with non-rotoscoped animation because of rotoscoping's increased realism. This does not mean that the technique is never used, in fact it was used almost exhaustively for Disney's 'Snow White' [THOMB76]. With modern three-dimensional computer animation this process is not sufficient and would feature high labour costs. The development of these techniques into the three-dimensional world developed what is now called motion capture.
Motion Capture, or MoCap for short also has many anachronisms within different paradigms, for example it is known as performance animation or digital motion, within other more traditional branches of animation. Motion capture was first introduced in the 1980's where animations were directly developed from human motion. Modern motion capture refers to computer-driven methods that let you hook up optical, magnetic or mechanical sensors to an actor's body, [BODE97, DYER95]. This allows the tracking of relative movement of the sensor nodes to reproduce a model of individualised human movement. The resultant information from a Motion Capture can then either be incorporated directly into an animation with highly realistic results or the data can be manipulated first to provide greater variety.
Editing and alteration of Motion Capture data is presently a very active research area in the realm of computer graphics. Most research concentrates upon techniques to get extra use from Motion capture data via combining motion captures, constraining movement in some way, spacetime-constraints or some other modification technique, [BRUD95, GLE97, GLE98, GRA99, MULT99, ROSE96, WIT95].
Even with the current advances made in adjusting Motion Capture data, it is a medium that still has limitations, mainly relating to the underlying restrictions whilst gathering original Motion Capture data. For example, it would be impractical to film a Motion Capture scene incorporating injury or destruction of the character, as this would entail the same effects occurring to the actor. The ability to create data for such events would be highly desirable for both the games industries and for medical science.
![]()
Title, Supervisors, Qualification aimed for, Computer Animation background, Animating the Human Form, Aims of the Project, Expected benefits to Industry, Project Plan, Required Resources, References, Menu Bar.
![]()