Equus:
The Virtual Horse
We are developing virtual horse models and animations. This project is part of Fauna, a more general research effort focussed on modeling and animating articulated creatures such as animals and humans.
Equus:
The Virtual Horse
| The Canonical Horse |
|
The anatomically-based "canonical" horse model is composed of multiple layers: structure hierarchy, bones, muscles, generalized tissues, skin, and coat. |
| Skinning |
|
Select "feature" vertices in the skin mesh are anchored to the underlying component (e.g. muscle, or bone), and move with the component when it is animated and/or deformed. The remaining vertices (the bulk of the mesh vertices) are interpolated based on the movement of nearby feature vertices. |
| 3D Model Morphing |
|
Given a relatively small set of measurements from a target horse, the canonical horse model is automatically morphed to generate a novel 3D model, complete with skin and animation hierarchy. |
| Motion Capture |
|
Using a combination of techniques from computer vision (such as active contours) and user interaction, horse motion can be extracted from monocular (single-camera) video. |
| Animations |
|
Some example animations demonstrating the skinning technique. The canonical horse is shown animated, as well as a pony. The pony model was automatically generated using our 3D model morphing technique from measurements taken from a live pony. |
| Rendering |
|
The images on this page were rendered using OpenGL with the default lighting and Gouraud shading. We are also working on a shader for sleek coats to generate more realistic imagery suitable for use with a ray-tracing based renderer. This is work in progress... more soon. |