THREE-DIMENSIONAL EYEGLASSES MODELING FROM TWO-DIMENSIONAL IMAGES

• US 20150362761A1
• Filed: 06/13/2015
• Published: 12/17/2015
• Est. Priority Date: 06/13/2014
• Status: Active Grant

• Current Assignee: eBay Inc.
• Original Assignee: eBay Inc.
• Inventors: Adler, Itay, Sincai, Dan
• Granted Patent: US 10,121,178 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 30/0621 Item configuration or customization
• 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 7/13 Edge detection
• CPC Class Codes: G06T 7/40 Analysis of texture depth or shape recovery from texture G06T7/529
• CPC Class Codes: G06T 7/55 from multiple images

Abstract

There are provided systems and method for three-dimensional (3D) eyeglasses modeling from two-dimensional (2D) images. A service provider, such as a payment provider, may access 2D images of a pair of eyeglasses, such as images from a front perspective and a side perspective of the eyeglasses. The contours of the eyeglasses may be extracted, such as the contours of the main features (e.g., the bridge, lens aperture, side and ear frame sections, and other important features). The contours may be processed to determine a curve that best fits the features. Additional features may be extracted, such as color, design, writing (e.g., a brand name on the frame), hinges, or other sub-feature of the frame. Utilizing this information, a 3D model of the pair of eyeglasses may be constructed by scaling the size of the lens curves, bridge, and/or arms and applying a texture.

Claims

1. A system comprising:

an eyeglasses modeling module comprising at least one hardware processor that accesses at least two two-dimensional (2D) images of a pair of eyeglasses, extracts contours of the pair of eyeglasses from the at least two 2D images, wherein the contours define shape features of the pair of eyeglasses, determines locations for the hinges of the pair of eyeglasses within the shape features using the at least two 2D images, and determines a three-dimensional (3D) model of the pair of eyeglasses using the contours and the hinges;

a database stored to a non-transitory memory that stores the at least two 2D images, the contours, the locations, and the 3D model; and

a network interface component that communicates the 3D model for display to a user.

2. The system of claim 1, wherein the contours are extracted by the eyeglasses modeling module using Canny edge detection.

3. The system of claim 1, wherein the eyeglasses modeling module further determines texture information for the pair of eyeglasses, wherein the 3D model is further determined using the texture information.

4. The system of claim 3, wherein the texture information is determined using at least one of the two 2D images of the pair of eyeglasses and information for the pair of eyeglasses.

5. The system of claim 4, wherein the information comprises at least one of material for the pair of eyeglasses, brand information for the pair of eyeglasses, weight for the pair of eyeglasses, lenses for the pair of eyeglasses, and color information for the pair of eyeglasses.

6. The system of claim 3, wherein the texture information comprises color, surface quality, surface properties, logos, shading, and lens.

7. The system of claim 1, wherein the contours further define curves of a lenses frame, a front, and arms for the pair of eyeglasses.

8. The system of claim 1, wherein the 3D model of the pair of eyeglasses is determined by generating a front, arms, and lenses of the pair of eyeglasses using the contours, and extruding the arms from the front.

9. The system of claim 1, wherein the eyeglasses modeling module further cleans up the contours to extract a best curve used to determine the 3D model.

10. The system of claim 1, wherein the eyeglasses modeling module determines the hinges using the contours.

11. The system of claim 7, wherein a meeting of two or more of the contours defines the hinges.

12. A method comprising:

receiving, by an eyeglasses modeling module comprising at least one hardware processor, at least two two-dimensional (2D) images of a pair of eyeglasses from a merchant device or server for a merchant, wherein the pair of eyeglasses is offered for sale by the merchant;

extracting, by the eyeglasses modeling module, contours of the pair of eyeglasses from the at least two 2D images, wherein the counters define a plurality of shape features of the pair of eyeglasses;

determining, by the eyeglasses modeling module, locations for the hinges of the pair of eyeglasses using the at least two 2D images;

determining, by the eyeglasses modeling module, a three-dimensional (3D) model of the pair of eyeglasses using the contours; and

communicating the 3D model to the merchant for display on a merchant website or mobile application for the merchant.

13. The method of claim 12, wherein the shape features comprise arms of the pair of eyeglasses, a frame of the pair of eyeglasses, and front of the pair of eyeglasses, and wherein the 3D model comprises the arms, the frame, and the front.

14. The method of claim 13, wherein a user accesses the merchant website using a communication device for the merchant, and wherein the user views the 3D model using the merchant website.

15. The method of claim 14, wherein the user communicates an image of the user to the merchant website, and wherein the merchant website displays the image of the user with the 3D model of the pair of eyeglasses.

16. The method of claim 15, wherein the arms are widened away from the frame and the front to fit a back of a head of the user in the image.

17. The method of claim 12, wherein the contours are extracted by the eyeglasses modeling module using Canny edge detection.

18. The method of claim 12, wherein the eyeglasses modeling module determines the hinges using the contours.

19. The method of claim 12, wherein prior to determining the 3D model, the method further comprises:

determining a nose bridge feature of the pair of eyeglasses using the contours,

wherein the 3D model is further determined using the nose bridge feature.

20. A non-transitory computer-readable medium comprising instructions which, in response to execution by a computer system, cause the computer system to perform a method comprising:

receiving, by an eyeglasses modeling module comprising at least one hardware processor, at least two two-dimensional (2D) images of a pair of eyeglasses from a merchant device or server for a merchant, wherein the pair of eyeglasses is offered for sale by the merchant;

extracting, by the eyeglasses modeling module, contours of the pair of eyeglasses from the at least two 2D images, wherein the counters define a plurality of shape features of the pair of eyeglasses;

determining, by the eyeglasses modeling module, locations for the hinges of the pair of eyeglasses using the at least two 2D images;

determining, by the eyeglasses modeling module, a three-dimensional (3D) model of the pair of eyeglasses using the contours; and

communicating the 3D model to the merchant for display on a merchant website or mobile application for the merchant.