TY - JOUR
T1 - Sample buffer
T2 - A progressive refinement ray-casting algorithm for volume rendering
AU - Ke, Hao Ren
AU - Chang, Ruei Chuan
N1 - Funding Information:
1. INTRODUCI'ION Recently, volume visualization has received a great deal of attention [ 1-5 ]. Initially, volume visualization was used to help researchers interpret medical images[l, 5 ] from CT, MRI, and other devices. Later, applications for volume visualization were found in other fields, such as fluid dynamics [ 3 ], earth science [ 2 ], and molecular chemistry[4] because experiments in these fields also generate volumetric data. There are several different approaches to rendering three-dimensional (3D) volumetric data. Some algorithms extract iso-surfaces from the volume and then display them[ 6-8 ]. Lorensen and Cline [ 8 ] proposed the marching cube technique. This method converts volume data into triangles by sequentially constructing iso-surfaces from a cube that is formed by eight closest voxels from two adjacent slices. Surface-extracting algorithms have their merits, as most graphic workstations can render polygons rapidly, but they also suffer from common shortcomings: information other than iso-surfaces cannot be explored and the iso-surfaces extracted are not necessarily correct. Methods that render volume data directly without transforming datasets into intermediate polygonal forms have also been discussed in the literature. These techniques are called volume rendering techniques [ 9-13]. They not only allow the intermediate transformation to be avoided, but also enable interior information to be observed. Some approaches [ 10, 13 ] project each cell of a volume into the image plane in either back-to-front or front-to-back order and then composite the projection of each cell into an accumulated image. Ray-casting is also used to render volume data directly[ 11, 12 ]. The input is a volume having f(X,) * This research was supported by the National Science Council of the Republic of China under grant No. NSC81-0408-E009-20.
PY - 1993/1/1
Y1 - 1993/1/1
N2 - Ray-casting has been shown to be an excellent technique for rendering volume data. This paper describes a new ray-casting algorithm that employs the concept of progressive refinement for volume rendering. Volume data are embedded into a complete octree, and when the volume data are rendered, the octree is traversed according to a user-defined threshold. By reducing this threshold gradually, we can render increasingly refined images. Ray-casting is used to explore the octree, and a data structure called a sample buffer (SB) is introduced to reduce the time required to render a series of images. With this method, a satisfactory balance between image quality and computation time can be achieved.
AB - Ray-casting has been shown to be an excellent technique for rendering volume data. This paper describes a new ray-casting algorithm that employs the concept of progressive refinement for volume rendering. Volume data are embedded into a complete octree, and when the volume data are rendered, the octree is traversed according to a user-defined threshold. By reducing this threshold gradually, we can render increasingly refined images. Ray-casting is used to explore the octree, and a data structure called a sample buffer (SB) is introduced to reduce the time required to render a series of images. With this method, a satisfactory balance between image quality and computation time can be achieved.
UR - http://www.scopus.com/inward/record.url?scp=0027590525&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0027590525&partnerID=8YFLogxK
U2 - 10.1016/0097-8493(93)90076-L
DO - 10.1016/0097-8493(93)90076-L
M3 - Article
AN - SCOPUS:0027590525
VL - 17
SP - 277
EP - 283
JO - Computers and Graphics
JF - Computers and Graphics
SN - 0097-8493
IS - 3
ER -