Abstract
We propose a new algorithm of automatic rendering for the volume data whose density is high for X-Y plane, but is low for Z, as in medical CT image data.
We detect rough wireframe outline data of objects from 3-D volume data, and trace the border voxels which are on each image plane. In this case, however, we cannot get smooth curved surface by rendering the voxels directly, because of their wider data's gap for the Z-axis direction. In order to solve this problem, we generate border data by cubic spline interpolation with adapting to the data of 3-D branch, confluence and multi-objects in automatic spline-point taking. In this way, we can have uniform density data points for X, Y and Z directions.
Creating small triangle patches for these data, and making surface rendering, we can get smooth curved surface.
