| OCR Text |
Show 6 and four values are accumulated: attenuated light intensity, maximum light value, distance of maximum value along the ray, and the centroid of voxel-value along the ray. Color is calculated in hue, saturation, and lightness tuples. Another approach to volume rendering is splatting. Splatting performs a frontto- back object-order traversal of the voxels in the volumetric dataset[41]. Each voxel's contribution to the image is calculated and composited using a series of table lookups. The first step in the splatting algorithms is to determine the order to traverse the volume. The face of the volume and the corner of the face closest to the image plane can be found by passing the coordinates of the corners of the volume through a viewing matrix. Voxels are splatted according to their distance from the image plane, with the closest voxel being splatted first. The voxels in one slice are splatted before the next slice is started. The value at each voxel is classified according to user specified color and opacity mapping functions. Thus we calculate the rgb tuple shaded according to the gradient and attenuated by the opacity mapping. The next step is to find the contribution of the tuple to an image buffer by projecting the voxel into the image space. When using orthographic projection, a circular reconstruction kernel is used to find the pixel extent of the contribution. Progressive refinement can be used to achieve interactive rates for rendering using the splatting technique[l9]. Another similar technique for volume rendering is the V-buffer algorithm[39]. It is a cell-based approach and uses a front-to-back object-order traversal and projection technique. The V-buffer technique traverses a many-point path within each cell, interpolating between cell corner values and projecting each interpolated value into an image buffer after the value has been color-mapped and shaded. The volume slice shearing algorithm described in Ref.[4] takes advantage of the unique hardware capabilities for the Pixar Image Computer. 2.2 Surface Fitting Algorithms In contour-connecting algorithms we take one slice at a time from the 3D data and trace out a closed contour in each slice. The contours in the adjacent slices are |
