Object separation for scanned assets

• US 10,366,470 B2
• Filed: 12/11/2017
• Issued: 07/30/2019
• Est. Priority Date: 12/11/2017
• Status: Active Grant

• Current Assignee: Hasivision, LLC
• Original Assignee: YouSpace, Inc.
• Inventors: Brunner, Ralph
• Primary Examiner(s): Armand, Marc Anthony
• Application Number:
  US15/838,228
• Publication Number:
  US 20190180410A1
• Time in Patent Office: 596 Days
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G06N 3/006 based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 30/0643 Graphical representation of items or shoppers
• CPC Class Codes: G06T 11/001 Texturing; Colouring; Generation of texture or colour inpainting G06T5/77
• CPC Class Codes: G06T 11/60 Editing figures and text; Combining figures or text
• CPC Class Codes: G06T 13/20 3D [Three Dimensional] animation
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06T 2210/16 Cloth
• CPC Class Codes: G06T 3/10 Selection of transformation methods according to the characteristics of the input images
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects

Abstract

Various of the disclosed embodiments present systems and methods for distinguishing portions of a virtual model associated with a clothing article from portions of the virtual model not associated with the clothing article. Some embodiments facilitate quick and effective separation by employing a feature vector structure conducive to separation by a linear classifier. Such efficient separation may be especially beneficial in applications requiring the rapid scanning of large quantities of clothing while retaining high-fidelity representations of the clothing'"'"'s geometry. Some embodiments further accommodate artist participation in the filtering process as well as scanning of articles from a variety of orientations and with a variety of supporting structures.

Claims

1. A computer system comprising:

at least one processor;

at least one interface configured to receive a scanned virtual model;

at least one memory comprising instructions configured to cause the system to perform a method separating a plurality of vertices associated with a clothing asset from a plurality of vertices associated with one or more objects other than the clothing asset, the method comprising;

receiving a first set of color values associated with the clothing asset;

receiving a second set of color values associated with a non-clothing object;

generating a first plurality of training feature vectors from the first set of color values;

generating a second plurality of training feature vectors from the second set of color values;

determining a plurality of feature component weights by providing the first plurality of feature vectors and the second plurality of feature vectors to a classifier;

receiving, via the at least one interface, the virtual model, the virtual model comprising;

a plurality of vertices, a first set of the plurality of vertices associated with clothing asset data and a second set of the plurality of vertices associated with non-clothing asset data; and

a map associating each vertex from the plurality of vertices with a corresponding color value; and

for a vertex of the plurality of vertices;

determining a color value associated with the vertex in the map;

generating a test feature vector based upon the color value;

applying the plurality of feature component weights to the test feature vector to determine a metric value; and

designating the vertex of the plurality of vertices for removal based upon the metric value.

2. The computer system of claim 1, wherein the first plurality of training feature vectors, the second plurality of training feature vectors, and the test feature vector comprise:

a first component color value;

a second component color value;

a third component color value;

a squared value of the first component color value;

a squared value of the second component color value;

a squared value of the third component color value;

a multiplied value of the first component color value and the third component color value;

a multiplied value of the second component color value and the third component color value; and

a multiplied value of the first component color value and the second component color value.

3. The computer system of claim 2, wherein the first component color value is a red value, the second component color value is a blue value, and the third component color value is a green value.

4. The computer system of claim 2, wherein the classifier is a support vector machine.

5. The computer system of claim 4, wherein

the support vector machine is a linear support vector machine, and wherein

applying the plurality of feature component weights to the test feature vector to determine a metric value comprises determining the dot product of the test feature vector with the plurality of feature component weights.

6. The computer system of claim 5, wherein the non-clothing object is a mannequin.

7. The computer system of claim 6, the method further comprising:

applying a rule from at least one of a plurality of system rules or a plurality of artist rules to the vertex designated for removal, the rule assessing color values associated with a plurality of vertices neighboring the vertex designated for removal; and

designating the vertex designated for removal to be retained based upon the application of the rule.

8. A non-transitory computer-readable medium comprising instructions configured to cause a computer system to perform a method comprising:

receiving a first set of color values associated with asset data;

receiving a second set of color values associated with non-asset data;

generating a first plurality of training feature vectors from the first set of color values;

generating a second plurality of training feature vectors from the second set of color values;

determining a plurality of feature component weights by providing the first plurality of feature vectors and the second plurality of feature vectors to a classifier;

receiving a virtual model comprising;

a plurality of vertices, a first set of the plurality of vertices associated with asset data and a second set of the plurality of vertices associated with non-asset data; and

a map associating each vertex from the plurality of vertices with a corresponding color value; and

for a vertex of the plurality of vertices;

determining a color value associated with the vertex in the map;

generating a test feature vector based upon the color value;

applying the plurality of feature component weights to the test feature vector to determine a metric value; and

designating the vertex of the plurality of vertices for removal based upon the metric value.

9. The non-transitory computer-readable medium of claim 8, wherein the first plurality of training feature vectors, the second plurality of training feature vectors, and the test feature vector comprise:

a first component color value;

a second component color value;

a third component color value;

a squared value of the first component color value;

a squared value of the second component color value;

a squared value of the third component color value;

a multiplied value of the first component color value and the third component color value;

a multiplied value of the second component color value and the third component color value; and

a multiplied value of the first component color value and the second component color value.

10. The non-transitory computer-readable medium of claim 9, wherein the first component color value is a red value, the second component color value is a blue value, and the third component color value is a green value.

11. The non-transitory computer-readable medium of claim 9, wherein the classifier is a support vector machine.

12. The non-transitory computer-readable medium of claim 11, wherein

the support vector machine is a linear support vector machine, and wherein

applying the plurality of feature component weights to the test feature vector to determine a metric value comprises determining the dot product of the test feature vector with the plurality of feature component weights.

13. The non-transitory computer-readable medium of claim 12, wherein the asset is an article of clothing and the non-asset is a mannequin.

14. The non-transitory computer-readable medium of claim 13, the method further comprising:

applying a rule from at least one of a plurality of system rules or a plurality of artist rules to the vertex designated for removal, the rule assessing color values associated with a plurality of vertices neighboring the vertex designated for removal; and

designating the vertex designed for removal to be retained based upon the application of the rule.

15. A computer-implemented method for separating asset data from non-asset data in a scanned virtual model, the method comprising:

receiving a first set of color values associated with asset data;

receiving a second set of color values associated with non-asset data;

generating a first plurality of training feature vectors from the first set of color values;

generating a second plurality of training feature vectors from the second set of color values;

determining a plurality of feature component weights by providing the first plurality of feature vectors and the second plurality of feature vectors to a classifier;

receiving a virtual model comprising;

a plurality of vertices, a first set of the plurality of vertices associated with asset data and a second set of the plurality of vertices associated with non-asset data; and

a map associating each vertex from the plurality of vertices with a corresponding color value; and

for a vertex of the plurality of vertices;

determining a color value associated with the vertex in the map;

generating a test feature vector based upon the color value;

applying the plurality of feature component weights to the test feature vector to determine a metric value; and

designating the vertex of the plurality of vertices for removal based upon the metric value.

16. The computer-implemented method of claim 15, wherein the first plurality of training feature vectors, the second plurality of training feature vectors, and the test feature vector comprise:

a first component color value;

a second component color value;

a third component color value;

a squared value of the first component color value;

a squared value of the second component color value;

a squared value of the third component color value;

a multiplied value of the first component color value and the third component color value;

a multiplied value of the second component color value and the third component color value; and

a multiplied value of the first component color value and the second component color value.

17. The computer-implemented method of claim 16, wherein the first component color value is a red value, the second component color value is a blue value, and the third component color value is a green value.

18. The computer-implemented method of claim 16, wherein the classifier is a support vector machine.

19. The computer-implemented method of claim 18, wherein

the support vector machine is a linear support vector machine, and wherein

applying the plurality of feature component weights to the test feature vector to determine a metric value comprises determining the dot product of the test feature vector with the plurality of feature component weights.

20. The computer-implemented method of claim 19, wherein the asset is an article of clothing and the non-asset is a mannequin.