Volume rendering of segmented tubular objects

Abstract

This paper describes a new method of combining ray casting with segmented tubular objects, such as blood vessels, for purposes of clinically useful display. The method first projects segmented tubes using a modified z-buffer that additionally records information about the objects projected. A subsequent step selectively volume renders only through the object volumes recorded by the z-buffer. In common with traditional "block" volume rendering the actual image data is shown, of importance when the boundary of a segmented object is uncertain. Unlike traditional "block" volume rendering, the approach permits user manipulation of objects, operates rapidly, and provides depth information even for maximum intensity projection. Although our methods were developed for display of the intracerebral vasculature, the approach is applicable to volume rendering of tubular objects throughout the body.