Parallel volume rendering system with a resampling module for parallel and perspective projections

• US 6,313,841 B1
• Filed: 04/13/1998
• Issued: 11/06/2001
• Est. Priority Date: 04/13/1998
• Status: Expired due to Term

• Current Assignee: TeraRecon Inc. (Fortive Corp.)
• Original Assignee: TeraRecon Inc. (Fortive Corp.)
• Inventors: Ogata, Masato, Ohkami, TakaHide, Lauer, Hugh C.
• Primary Examiner(s): Vo, Cliff N.
• Application Number:
  US09/059,155
• Time in Patent Office: 1,303 Days
• Field of Search: 345/419, 345/424, 345/427, 345/440, 345/502, 345/505
• US Class Current: 345/424
• CPC Class Codes: G06T 15/10 Geometric effects

Abstract

A volume rendering system re-samples voxels read from a voxel memory to generate samples along perspective rays cast from a center of projection using a level of detail value. Color computations are performed with the samples to produce pixels for a baseplane image. The level of detail is computed, at each plane of samples perpendicular to a principal viewing axis, from the current sample position and the distance between the center of projection and the baseplane; the principal viewing axis is the coordinate axis in a rendered volume most parallel with a viewing vector. The level of detail provides a measure of the distance between two neighboring perspective rays at each plane and is used to determine the number of voxels and weights for these voxels required to compute a single sample at each plane. Multi-resolution datasets prepared for different levels of details are used to simplify the resampling operation by limiting the number of voxels required to compute a single sample.

Claims

1. A volume rendering system for enabling parallel-pipeline processing of perspective rays cast into a volume to be rendered by parallizing perspective rays, comprising:

a memory for storing plural voxels, said voxels collectively describing said volume to be rendered;

a resampling unit for selectively receiving a subset of said voxels from said memory, said subset of said voxels proximate to a sample along a particular ray, and for combining said subset of voxels into said sample;

a rendering module for receiving said sample and for selectively combining said sample with other samples along said particular ray received from said resampling unit to generate a single pixel centered on said particular ray; and

a perspective ray control element, in communication with said memory, said resampling unit, and said rendering module, for casting said particular ray through said single pixel.

2. The volume rendering system of claim 1, wherein said perspective ray control is for identifying which of said plural voxels are to comprise said subset of said voxels.

3. The volume rendering system of claim 1, wherein said perspective ray control is for specifying to said resampling unit a weight to be applied to each voxel of said subset of voxels.

4. The volume rendering system of claim 1, wherein said perspective ray control is for specifying to said resampling unit a range of voxels of said subset of voxels to be combined in the generation of said sample.

5. The volume rendering system of claim 1, wherein said rendering module further comprises a gradient estimation unit, a shading unit, and a compositing unit for compositing said samples.

6. The volume rendering system of claim 5, wherein said perspective ray control is for adjusting said gradient estimation and shading units.

7. The volume rendering system of claim 1, further comprising a baseplane image memory for storing said pixel and other pixels received from said rendering unit.

8. The volume rendering system of claim 1, wherein said perspective ray control is further for calculating a level of detail for said sample, said level of detail effecting a weight to be applied to each voxel of said subset of voxels by said resampling unit and a range of voxels of said subset of voxels to be combined by said resampling unit.

9. The volume rendering system of claim 8, wherein said perspective ray control is further for calculating said level of detail based upon a distance from a center of projection for plural perspective rays to a baseplane and on a current position of said sample generated by said resampling unit.

10. A volume rendering system for enabling parallel-pipeline processing of perspective rays cast into a volume to be rendered by parallelizing perspective rays, comprising:

a voxel memory for storing plural voxels collectively describing said volume to be rendered;

a dataset memory for storing plural multi-resolution datasets;

a resampling unit for selectively receiving a subset of said multi-resolution datasets from said memory, said subset proximate to a sample along a particular ray, and for combining said subset into a sample;

a rendering module for receiving said sample and for combining said sample with other samples along said particular ray received from said resampling unit to generate a single pixel;

a perspective ray control element, in communication with said dataset memory, said resampling unit, and said rendering module, for casting said particular ray through said single pixel; and

a processor in communication with said voxel memory and said dataset memory,

wherein said processor calculates said multi-resolution datasets based upon said voxels, such that said multi-resolution datasets represent each position within said volume to be rendered at varying degrees of resolution.

11. The volume rendering module of claim 10, wherein said perspective ray control is for identifying which of said plural multi-resolution datasets are to comprise said subset of said multi-resolution datasets.

12. The volume rendering system of claim 10, wherein said perspective ray control is for specifying to said resampling unit a weight to be applied to each multi-resolution dataset of said subset of multi-resolution datasets.

13. The volume rendering system of claim 10, wherein said perspective ray control is for specifying to said resampling unit a range of multi-resolution datasets of said subset of multi-resolution datasets which are to be combined in the generation of said sample.

14. The volume rendering system of claim 10, wherein said rendering module further comprises a gradient estimation unit, a shading unit, and a compositing unit for compositing said samples.

15. The volume rendering system of claim 14, wherein said perspective ray control is for adjusting said gradient estimation and shading units.

16. The volume rendering system of claim 10, further comprising a baseplane image memory for storing said pixel and other pixels received from said rendering unit.

17. The volume rendering system of claim 10, wherein said perspective ray control is further for calculating a level of detail for said sample, said level of detail effecting which of said multi-resolution datasets are to comprise said subset of multi-resolution datasets.

18. The volume rendering system of claim 17, wherein said perspective ray control is further for calculating said level of detail based upon a distance from a center of projection for plural perspective ray s to a baseplane and on a current position of said sample generated by said resampling unit.

19. A method of rendering a multi-dimensional volume on a baseplane via parallel-pipeline processing of perspective rays cast into said volume as parallelized perspective rays, comprising the steps of:

storing a representation of said volume as a plurality of voxels in a voxel memory;

retrieving a subset of said voxels from said memory by a resampling unit, said subset of said voxels proximate to a sample along a particular ray;

determining a range of said subset of voxels to be processed by said resampling unit;

determining a weight to be applied to each of said subset of voxels to be processed by said resampling unit;

compositing said weighted voxels within said range of voxels, by said resampling unit, to generate said sample; and

rendering a pixel from said sample and from other samples along said particular ray received from said resampling unit by a rendering module.

20. The method of claim 19, further comprising the step of calculating a level of detail for a sample to be calculated in a perspective ray control unit, and providing said level of detail to said resampling unit for use in said steps of determining a range and determining a weight.

21. The method of claim 20, wherein said step of calculating is preceded by the steps of:

calculating a distance from a center of projection for a set of perspective rays cast into said volume; and

calculating the relative position of a sample to be calculated within said volume,

wherein said distance and relative position are utilized in said step of calculating a level of detail.

22. A method of rendering a multi-dimensional volume on a baseplane via parallel-pipeline processing of perspective rays cast into said volume as parallelized perspective rays, comprising the steps of:

storing a representation of said volume as a plurality of voxels in a voxel memory;

generating a set of multi-resolution datasets from said voxels, said multi-resolution datasets representing each position within said volume at varying degrees of resolution,

retrieving a subset of said multi-resolution datasets from said memory by a resampling unit;

determining a range of said subset of said multi-resolution datasets to be processed by said resampling unit;

determining a weight to be applied to each of said subset of said multi-resolution datasets to be processed by said resampling unit;

compositing said weighted multi-resolution datasets within said range of multi-resolution datasets, by said resampling unit, to generate a sample; and

rendering a pixel from said sample and from other samples received from said resampling unit by a rendering module.

23. The method of claim 22, further comprising the step of calculating a level of detail for a sample to be calculated in a perspective ray control unit, and providing said level of detail to said resampling unit for use in said steps of determining a range and determining a weight.

24. The method of claim 23, wherein said step of calculating is preceded by the steps of:

calculating a distance from a center of projection for a set of perspective rays cast into said volume; and

calculating the relative position of a sample to be calculated within said volume,

wherein said distance and relative position are utilized in said step of calculating a level of detail.

25. A method for rendering voxels onto a baseplane including a plurality of pixels using parallel processing pipelines, comprising the steps of:

casting a perspective ray, for each pixel of the baseplane, through the voxels, each perspective ray including a plurality of sample points;

sampling voxels proximate to each sample point along a particular ray associated with a particular pixel in the parallel processing pipelines; and

accumulating the sampled voxels of each sample point along the particular ray into the particular pixel of the baseplane to perspectively render the voxels using the parallel processing pipelines.

26. The method of claim 25 further comprising the step of:

weighting the sampled voxels according to a proximity to each sample point in parallel.