Image rendering utilizing procedural yarn model generated in multi-stage processing pipeline

Patent metadata

Field	Value
Publication number	US-10410380-B2
Application number	US-201715636213-A
Country	US
Kind code	B2
Filing date	Jun 28, 2017
Priority date	Jun 29, 2016
Publication date	Sep 10, 2019
Grant date	Sep 10, 2019

Abstract

An apparatus in one embodiment comprises a multi-stage processing pipeline configured to generate a procedural yarn model by fitting procedural yarn model parameters to input data comprising computed tomography measurements of one or more actual yarn samples. The apparatus further comprises an image rendering system configured to execute one or more procedural yarn generation algorithms utilizing the procedural yarn model and to generate based at least in part on results of execution of the one or more procedural yarn generation algorithms at least one corresponding output image for presentation on a display. The multi-stage processing pipeline comprises a first stage configured to perform ply twisting estimation and ply cross-sectional estimation, a second stage configured to classify constituent fibers into regular fibers and flyaway fibers, and third and fourth stages configured to process the respective regular and flyaway fibers to fit respective different sets of parameters of the procedural yarn model.

First claim

What is claimed is:

1. An apparatus for model-based generation of one or more images, comprising:
a multi-stage processing pipeline configured to generate a procedural yarn model by fitting procedural yarn model parameters to input data comprising computed tomography measurements of one or more actual yarn samples; and
an image rendering system configured to execute one or more procedural yarn generation algorithms utilizing the procedural yarn model and to generate based at least in part on results of execution of the one or more procedural yarn generation algorithms at least one corresponding output image for presentation on a display;
the multi-stage processing pipeline comprising:
a first stage configured to perform ply twisting estimation and ply cross-sectional estimation on at least a portion of the input data corresponding to a given yarn sample having multiple twisted plies with each such ply comprising multiple fibers;
a second stage configured to classify constituent fibers of each of the plies of the given yarn sample characterized by the input data into regular fibers of the input data and flyaway fibers of the input data;
a third stage configured to process the regular fibers; and
a fourth stage configured to process the flyaway fibers;
wherein the third and fourth stages fit different sets of parameters of the procedural yarn model using the respective regular and flyaway fibers;
the multi-stage processing pipeline and the image rendering system being implemented on at least one processing device comprising a processor coupled to a memory.

2. The apparatus of claim 1 wherein the input data is generated by a micro computed tomography scanner and comprises volumetric density information for each of multiple scanned yarn samples.

3. The apparatus of claim 1 wherein the input data for the given yarn sample is preprocessed prior to application to the multi-stage processing pipeline in order to generate extracted fiber curves and a plurality of tracked ply centers.

4. The apparatus of claim 1 wherein the first stage of the multi-stage processing pipeline is configured to fit additional parameters of the procedural yarn model to the input data and wherein the additional parameters comprise ply twisting parameters, ply cross-section parameters and per-ply fiber count.

5. The apparatus of claim 4 wherein the first stage of the multi-stage processing pipeline is further configured to utilize the additional parameters fit to the input data in that stage to adjust component plies of the given yarn sample so that all of the plies are centered on a common axis and have common cross-sectional shapes.

6. The apparatus of claim 4 wherein the ply twisting parameters are fit to the input data in the first stage of the multi-stage processing pipeline by computing a circularly helical curve for each of the plies as a function of a common ply radius shared by all the plies, a common ply pitch shared by all the plies, and a ply initial angle of a first one of the plies which is utilized to compute initial ply angles for other ones of the plies.

7. The apparatus of claim 6 wherein the common ply radius, the common ply pitch and the ply initial angle of a first one of the plies are determined by minimizing distance between a given one of the circularly helical curves and a corresponding one of a plurality of tracked ply centers of the input data.

8. The apparatus of claim 4 wherein the ply cross-section parameters are fit to the input data in the first stage of the multi-stage processing pipeline by:
for each of a plurality of cross-sectional planes, placing an ellipse at a center of each ply within that plane with a short axis of the ellipse pointing towards a center of a polygon formed by all of the ply centers and then transforming each ply in the plane to make it axis-aligned and centered at a common origin;
combining the transformed plies from each of the plurality of cross-sectional planes to obtain an accumulated two-dimensional point cloud; and
computing separate ply cross-section parameters for x and y dimensions as a function of respective standard deviations of x and y coordinates of points in the point cloud.

9. The apparatus of claim 1 wherein the second stage of the multi-stage processing pipeline is configured to classify constituent fibers of each of the plies into regular fibers and flyaway fibers by:
determining for each of a plurality of fibers a minimal and maximal distance of that fiber from its corresponding ply center;
eliminating any fibers having minimal distances above a first threshold; and
for any remaining fibers, classifying those fibers with maximal distances at or below a second threshold as regular fibers and those with maximal distances above the second threshold as flyaway fibers.

10. The apparatus of claim 1 wherein the parameters of the procedural yarn model that are fit utilizing regular fibers of the input data in the third stage of the multi-stage processing pipeline comprise cross-sectional fiber distribution parameters, fiber twisting parameters and fiber migration parameters.

11. The apparatus of claim 10 wherein the fiber migration parameters are determined by minimizing reconstruction error between the regular fibers and respective corresponding generated fibers over all of the regular fibers.

12. The apparatus of claim 11 wherein the cross-sectional fiber distribution parameters are determined from the fiber migration parameters by per-fiber error minimization of at least a subset of the fiber twisting parameters followed by maximum likelihood estimation over results of the per-fiber error minimization.

13. The apparatus of claim 1 wherein the parameters of the procedural yarn model that are fit utilizing flyaway fibers of the input data in the fourth stage of the multi-stage processing pipeline comprise flyaway fiber distribution parameters of a flyaway fiber model.

14. The apparatus of claim 13 wherein the flyaway fiber model characterizes the flyaway fibers as comprising loop-type flyaway fibers and hair-type flyaway fibers.

15. The apparatus of claim 14 wherein the flyaway fiber distribution parameters for a given one of the loop-type fibers comprises a radius scaling parameter for application to selected vertices characterizing respective distances to a ply center and a loop-type fiber density parameter that controls a number of applications of the radius scaling parameter.

16. The apparatus of claim 14 wherein the flyaway fiber distribution parameters for a given one of the hair-type fibers comprises starting and ending radii parameters characterizing respective distances to a ply center, corresponding azimuthal angles and locations that in combination with the starting and ending radii parameters specify an arc shape, and a hair-type fiber density parameter that controls a number of arc shapes to be generated.

17. The apparatus of claim 14 wherein at least one of the flyaway fibers contains both loop-type and hair-type components and is separated into corresponding respective segments for determination of the flyaway fiber distribution parameters.

18. The apparatus of claim 1 wherein at least the third and fourth stages of the multi-stage processing pipeline are configured to operate at least in part in parallel with one another.

19. The apparatus of claim 1 wherein a given one of the one or more procedural yarn generation algorithms utilizing the procedural yarn model comprises a realization-reducing procedural yarn generation algorithm which is configured to determine l

Assignees

• Univ Cornell
• Univ California

Inventors

• Bala Kavita
• Luan fujun
• Zhao Shuang

Classifications

• G06T12/00

  Tomographic reconstruction from projections · CPC title

• G06T12/30 Primary

  Image post-processing, e.g. metal artefact correction · CPC title

• G06F18/24

  Classification techniques · CPC title

• G06T15/00

  Three-dimensional [3D] image rendering · CPC title

• G06T2210/52

  Parallel processing · CPC title