Modeling anisotropic surface reflectance with microfacet synthesis

Patent metadata

Field	Value
Publication number	US-9098945-B2
Application number	US-43391009-A
Country	US
Kind code	B2
Filing date	May 1, 2009
Priority date	May 1, 2009
Publication date	Aug 4, 2015
Grant date	Aug 4, 2015

Abstract

Described is a search technology in which spatially varying anisotropic reflectance is modeled using image data captured from a single view. Reflectance at each point is represented using a microfacet-based Bidirectional Reflectance Distribution Function (BRDF). Modeling processes the image data, which provides a partial normal distribution function (NDF) for each surface point. The NDF at each selected point is completed by texture synthesis using similar, overlapping partial NDFs from other points. Also described is a scanning device that illuminates a sample surface from a two-dimensional set of light directions using a linear array of LEDs moved over a flat sample.

First claim

What is claimed is:

1. In a computing environment, a method comprising, processing a set of images associated with a sample surface into partial normal distribution functions corresponding to points of the sample surface, each image captured from a different illumination position with respect to illuminating the sample surface, completing the partial normal distribution functions into completed normal distribution functions for at least some selected partial normal distribution functions corresponding to at least some selected points of the sample surface, including by synthesizing each partial normal distribution corresponding to a different selected point using overlapping partial normal distribution functions of other points of the sample surface, and using the completed normal distribution functions to obtain bidirectional reflectance distribution functions representing reflectance information at a plurality of points of the sample surface.

2. The method of claim 1 further comprising, performing clustering to select which partial normal distribution functions to complete.

3. The method of claim 1 wherein for each selected partial normal distribution function, the other partial normal distribution functions correspond to a set of search candidates, and further comprising, pruning the search candidates into a smaller subset of search candidates.

4. The method of claim 1 wherein for each selected partial normal distribution function, the other partial normal distribution functions correspond to a search space, and further comprising, pruning the search space based upon how much the partial normal distribution function domains overlap, and how similar the normal distribution functions are over that overlapping domain.

5. The method of claim 1 wherein for each selected partial normal distribution function, the other partial normal distribution functions correspond to a search space, and further comprising, extending the search space using synthesis operations on initially measured data of the normal distribution functions.

6. The method of claim 1 wherein using the completed normal distribution functions to obtain the bidirectional reflectance distribution functions further comprises normalizing each completed normal distribution function.

7. The method of claim 1 wherein using the completed normal distribution functions to obtain the bidirectional reflectance distribution functions further comprises computing a shadowing term for each selected point based upon the completed normal distribution function corresponding to that point.

8. The method of claim 5 further comprising estimating a relative refraction index for each selected point.

9. The method of claim 1 further comprising, changing each illumination position to a different illumination position by: (a) setting a linear array of light sources at a start position; (b) individually illuminating light sources of the linear array; (c) moving the linear array towards a stop position in a direction that is substantially perpendicular to the linear array; and (d) repeating from step (b) until the stop position is reached.

10. The method of claim 1 further comprising, capturing the images corresponding to the different illumination positions with a single camera that is fixed with respect to the different illumination positions.

11. The method of claim 1 further comprising, outputting spatially varying bidirectional reflectance distribution functions corresponding to the bidirectional reflectance distribution functions for use in rendering computer graphics.

12. In a computing environment, a system comprising, a scan mechanism configured to illuminate a sample surface from a plurality of illumination positions in two dimensions via a set of individually illuminated light sources, the scan mechanism coupled to a camera that is configured to capture, from at least one fixed view direction, a set of images corresponding to the illumination positions, the illumination positions comprise positions where the set of individually illuminated light sources does not occlude the camera, the images provide overlapping partial normal distribution functions that are processed into normal distribution functions for selected points of the sample surface to determine a bidirectional reflectance distribution function at each selected point, the overlapping partial normal distribution functions corresponding to other points of the sample surface.

13. The system of claim 12 wherein the set of individually illuminated light sources scan mechanism comprises a linear array of LEDs.

14. The system of claim 13 wherein the scan mechanism comprises a motor configured to move the linear array of LEDs in a direction that is perpendicular to the linear array.

15. The system of claim 14 wherein the motor is further configured to move the linear array towards or away from the camera.

16. The system of claim 12 wherein the modeling process is configured to compute the normal distribution function for each selected point based upon the partial normal distribution function corresponding to that point and other partial normal distribution functions corresponding to other points.

17. The system of claim 12 wherein the modeling process includes a clustering mechanism by which the selected points are selected.

18. One or more computer-readable storage media having computer-executable instructions not consisting of a signal, which when executed perform steps, comprising:
processing a set of images associated with a sample surface into partial normal distribution functions corresponding to points of the sample surface, each image being captured from a single viewing direction with respect to a different illumination position for illuminating the sample surface, including changing to the different illumination position by, (a) setting a linear array of light sources at a start position, (b) individually illuminating light sources of the linear array, (c) moving the linear array towards a stop position, and (d) repeating from step (b) until the stop position is reached;
modifying reflectance data associated with a grazing angle;
selecting a set of partial normal distribution functions, each selected partial normal distribution function corresponding to a selected point of the sample surface;
and for each selected partial normal distribution function, choosing search candidates corresponding to other partial normal distribution functions of other points of the sample surface, and completing that partial normal distribution function into a completed normal distribution function by applying a texture synthesis using the search candidates.

19. The one or more computer-readable storage media of claim 18 having further computer-executable instructions comprising, using the completed normal distribution functions to obtain bidirectional reflectance distribution functions, including, for each selected point, computing a specular coefficient based upon the completed normal distribution function, computing a shadowing term based upon the completed normal distribution function corresponding to that point, and estimating a relative refraction index.

20. The one or more computer-readable storage media of claim 18 having further computer-executable instructions comprising, synthesizing each partial normal distribution corresponding to a different selected point using overlapping partial normal distribution functions of the other points of the sample surface, and using the completed normal distribution functions to obtain bidir

Assignees

• Wang Jiaping
• Zhao Shuang
• Tong Xin
• Snyder John M
• Guo Baining
• Microsoft Technology Licensing Llc

Inventors

• Wang Jiaping
• Zhao Shuang
• Tong Xin
• Snyder John M
• Guo Baining

Classifications

• G06T15/506 Primary

  Illumination models · CPC title