Ray tracing apparatus and method

1. An image processing apparatus, comprising:

a processor;

a splitting unit, executing on the processor, to split a first space within an input three-dimensional (3D) model into a plurality of subspaces in order to generate an acceleration structure of the input 3D model;

a decision unit, executing on the processor, to set a subspace determined as having a relatively high probability of including a ray progress path among the plurality of subspaces, as a child node having a relatively high traversal priority in the acceleration structure among a plurality of child nodes; and

a flag setting unit, executing on the processor, to add, to the child node having the relatively high traversal priority, a flag bit used to identify whether a subspace relatively close to an origin in a world coordinate system among the plurality of subspaces is matched with the subspace determined to have the relatively high probability of including the ray progress path,

wherein the child node having the relatively high traversal priority is arranged in a same caching block to a parent node of the child node having the relatively high traversal priority based on a depth-first layout (DFL).

2. The image processing apparatus of claim 1 , wherein the acceleration structure corresponds to one data structure of a kd-tree, a bounding volume hierarch (BVH), and a grid.

3. The image processing apparatus of claim 1 , wherein the child node having the relatively high traversal priority corresponds to a left child node that is traversed in a depth first tree traversal prior to other child nodes.

4. The image processing apparatus of claim 1 , wherein the splitting unit designates a split axis for splitting the first space, sets at least one split surface candidate by aligning a plurality of primitives within the first space along the split axis, selects a split surface from the at least one split surface candidate according to a surface area heuristic based algorithm, and splits the first space into the plurality of subspaces using the selected split surface.

5. The image processing apparatus of claim 1 , wherein the splitting unit recursively repeats the splitting process with respect to a space non-split within the input 3D model until a leaf node generation criterion is satisfied yielding a plurality of children of the child node.

6. The image processing apparatus of claim 1 , further comprising:

a memory to store the generated acceleration structure; and

a ray tracer to render an image observed at a predetermined viewpoint according to a ray tracing scheme, using at least one of the stored acceleration structure.

7. The image processing apparatus of claim 1 , wherein the child node determined as having the relatively high probability of including the ray among the plurality of child nodes is set by the decision unit as a left child node that is traversed in a depth first tree traversal prior to the plurality of child nodes and a child node determined as having a relatively low probability of including the ray among the plurality of child nodes is set by the decision unit as a right child node.

8. The image processing apparatus of claim 1 , wherein the decision unit determines a spatial structure determined as having a relatively large surface area among the plurality of subspaces, as the subspace having the relatively high probability of including the ray progress path.

9. An image processing method, comprising:

splitting, by way of a processor, a first space within an input three-dimensional (3D) model into a plurality of subspaces in order to generate an acceleration structure of the input 3D model;

setting a subspace determined as having a relatively high probability of including a ray progress path among the plurality of subspaces, as a child node having a relatively high traversal priority in the acceleration structure among a plurality of child nodes; and

adding, to the child node having the relatively high traversal priority, a flag bit used to identify whether a subspace relatively close to an origin in a world coordinate system among the plurality of subspaces is matched with the subspace determined to have the relatively high probability of including the ray progress path,

wherein the child node having the relatively high traversal priority is arranged in a same caching block to a parent node of the child node having the relatively high traversal priority based on a depth-first layout (DFL).

10. The method of claim 9 , wherein the acceleration structure corresponds to one data structure of a kd-tree, a bounding volume hierarch (BVH), and a grid.

11. The method of claim 9 , wherein the child node having the relatively high traversal priority corresponds to a left child node that is traversed in a depth first tree traversal prior to other child nodes.

12. The method of claim 9 , wherein the splitting comprises:

designating a split axis for splitting the first space;

setting at least one split surface candidate by aligning a plurality of primitives within the first space along the split axis;

selecting a split surface from the at least one split surface candidate according to a surface area heuristic based algorithm; and

splitting the first space into the plurality of subspaces using the selected split surface.

13. The method of claim 9 , wherein the splitting is recursively repeated until a leaf node generation criterion is satisfied yielding a plurality of children of the child node.

14. The method of claim 9 , further comprising:

storing the generated acceleration structure; and

rendering an image observed at a predetermined viewpoint according to a ray tracing scheme, using at least one of the stored acceleration structure.

15. The method of claim 9 , wherein the child node determined as having the relatively high probability of including the ray among the plurality of child nodes is set as a left child node that is traversed in a depth first tree traversal prior to the plurality of child nodes and a child node determined as having a relatively low probability of including the ray among the plurality of child nodes is set as a right child node.

16. The method of claim 9 , wherein the setting comprises determining a spatial structure determined as having a relatively large surface area among the plurality of subspaces, as the subspace having the relatively high probability of including the ray progress path.

17. A non-transitory computer-readable medium comprising a program for instructing a computer to perform the method of claim 9 .