Methods and systems for acquiring svBRDF measurements

1. A method of obtaining a spatially varying bidirectional reflectance distribution function (svBRDF) for a flat object, comprising:

using a first video capture device and a first light source to capture a first flash video of an area of interest at a relatively fixed distance less than 50 cm over the surface of the flat object;

aligning video frame images from the captured first flash video into a single panorama with observations from multiple light and view locations for each of a plurality of pixels;

reviewing each of the video frame images from the captured first flash video and removing blurry video frames detected upon review of each of the video frame images from the captured first flash video;

obtaining a connected sub-graph of a connectivity map of video frames that remain after removing the blurry video frames;

removing video frames not contained in the connected sub-graph of the connectivity map;

clustering the plurality of pixels into a plurality of clusters by similar appearance; and

fitting a bidirectional reflectance distribution function (BRDF) to each of the plurality of clusters to generate a high-resolution svBRDF.

2. The method of claim 1 , further comprising:

placing a plurality of fiducial markers around an area of interest of the flat object; and

using the plurality of fiducial markers to establish physical locations.

3. The method of claim 2 , wherein the placing includes placing each of a plurality of printed fiducial markers at an edge of an area of interest.

4. The method of claim 3 , further comprising obtaining a single reference image of the area of interest based on the plurality of printed fiducial markers.

5. The method of claim 4 , further comprising establishing a global coordinate system based on the single reference image.

6. The method of claim 1 , further comprising recursively subdividing and fitting a BRDF and normal vector displacement to each sub-cluster.

7. The method of claim 6 , wherein the recursive subdividing and fitting continue until a residual fit error at each pixel is sufficiently low.

8. The method of claim 1 , further comprising applying the high-resolution svBRDF to a three-dimensional (3D) object for rendering in an application.

9. The method of claim 1 , wherein the first video capture device is a mobile phone.

10. The method of claim 1 , wherein the first video capture device is coupled with the first light source.

11. The method of claim 1 , further comprising using a second video capture device and a second light source to capture a second flash video of the area of interest at a relatively fixed distance over the surface of the flat object.

12. The method of claim 11 , further comprising aligning video frame images from the captured second flash video into the single panorama.

13. The method of claim 11 , wherein the second video capture device and the second light source are separate from each other.

14. The method of claim 11 , wherein the first and second light sources are the same light source.

15. The method of claim 1 , further comprising:

computing a shifted and scaled Fourier transform for each of the video frame images to detect the blurry video frames that are rejected.

16. The method of claim 1 , wherein the using includes moving the first video capture device over the area of interest in a systematic manner.

17. The method of claim 16 , wherein the systematic manner includes sinusoidal top-to-bottom and side-to-side manner.

18. The method of claim 1 , further comprising carrying out a feature matching process on the first flash video, and removing video frames of the first flash video due to insufficient feature matches prior to obtaining the connected sub-graph.

19. The method of claim 1 , wherein the connectivity map is a matrix of inlier counts, and wherein the connected sub-graph of the connectivity map has the most members.

20. A system for obtaining a spatially varying bidirectional reflectance distribution function (svBRDF) for a flat object, comprising:

at least one light source to provide light for a flash video of an area of interest at a relatively fixed distance less than 50 cm over the surface of the flat object;

at least one video capture device configured to capture the flash video; and

a processor configured to:

align video frame images from the captured flash video into a single panorama with observations from multiple light and view locations for each of a plurality of pixels;

review each of the video frame images as originally captured from the captured flash video and remove blurry video frames of the video frame images reviewed from the captured first flash video as originally captured;

obtain a connected sub-graph of a connectivity map of video frames that remain after removing the blurry video frames;

remove video frames not contained in the connected sub-graph of the connectivity map;

cluster the plurality of pixels into a plurality of clusters by similar appearance; and

fit a bidirectional reflectance distribution function (BRDF) to each of the plurality of clusters to generate a high-resolution svBRDF.