The proposed raytracer allows to create simple synthetic scenes and render them from different camera positions. Some properties of the objects (position, orientation, size) can be defined as constants or as functions of time. The actual engine supports texture mapping and antialiasing (shadows, reflections and refractions are not supported yet).
The raytracer might be useful for researchers working in the computer vision domain, because it not only renders images but also can compute ray hit point's absolute and relative (wrt objects) positions and indices (which face of which object has been hit). This feature allows estimation of disparity (horizontal and vertical) map and motion field.
Few videos, demonstrating some properties of the raytracer can be found on my web-site: http://sites.google.com/site/chumerin/projects/myraytracer