Computer vision assisted item search

• US 9,940,728 B2
• Filed: 12/15/2015
• Issued: 04/10/2018
• Est. Priority Date: 12/15/2015
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Baca, Jim S, Stanasolovich, David
• Primary Examiner(s): NAKHJAVAN, SHERVIN K
• Application Number:
  US14/970,023
• Publication Number:
  US 20170169580A1
• Time in Patent Office: 847 Days
• Field of Search: None
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2207/20024 Filtering details
• CPC Class Codes: G06T 2207/20221 Image fusion; Image merging
• CPC Class Codes: G06T 7/40 Analysis of texture depth or shape recovery from texture G06T7/529
• CPC Class Codes: G06T 7/50 Depth or shape recovery
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G06V 10/54 relating to texture
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Abstract

System and techniques for computer vision assisted item search are described herein. A composite image, including visual data and depth data, may be obtained. The composite image may be filtered to isolate a clothing article represented in the composite image. A classifier may be applied to the depth data to produce a set of clothing attributes for the clothing article. The clothing attributes may then be provided to a remote device.

Claims

1. A system for computer vision assisted item search, the system comprising:

an interface to obtain a composite image from a sensor, the composite image including visual data and depth data;

a filter to;

isolate a clothing article represented in the composite image; and

determine a clothing type of the clothing article based on the depth data;

a classifier to;

accept the depth data; and

produce a set of clothing attributes that correspond to the clothing article based on the depth data and the clothing type; and

a transceiver to provide the set of clothing attributes to a remote device.

2. The system of claim 1, wherein the set of clothing attributes include at least one of a size, a shape, or a texture.

3. The system of claim 2, wherein the classifier is to:

determine a texture of a fabric using the depth data; and

identify a material of the fabric that is correlated to the texture, and wherein the texture and material are included in the set of clothing attributes.

4. The system of claim 2, wherein the classifier is to:

scale a representation of the clothing article in the composite image using the depth data to produce an estimate of absolute measurements of the clothing article; and

convert the estimate of absolute measurements to a clothing article size.

5. The system of claim 2, wherein the classifier is to directly convert the depth data into a clothing article size.

6. The system of claim 1, comprising the remote device, the remote device to update a user profile with the clothing attributes, the clothing attributes corresponding to a clothing item object in a data structure of the user profile.

7. The system of claim 6, comprising a search engine, the search engine to search for additional clothing items using the clothing attributes and the user profile.

8. The system of claim 7, wherein to search for additional clothing items includes the search engine to apply a restriction that the additional clothing items differ from the clothing article in at least one of color, pattern, style, or type.

9. A method for computer vision assisted item search, the method comprising:

obtaining a composite image from a sensor array, the composite image including visual data and depth data;

filtering the composite image to isolate a clothing article represented in the composite image and determining a clothing type of the clothing article based on the depth data;

applying a classifier to the depth data to produce a set of clothing attributes that correspond to the clothing article based on the depth data and clothing type; and

providing the set of clothing attributes to a remote device.

10. The method of claim 9, wherein the set of clothing attributes include at least one of a size, a shape, or a texture.

11. The method of claim 10, wherein applying the classifier to the depth data to produce the set of clothing attributes includes:

determining a texture of a fabric using the depth data; and

identifying a material of the fabric that is correlated to the texture, and wherein the texture and material are included in the set of clothing attributes.

12. The method of claim 10, wherein the classifier scales a representation of the clothing article in the composite image using the depth data to produce an estimate of absolute measurements of the clothing article, and wherein the classifier converts the estimate of absolute measurements to a clothing article size.

13. The method of claim 10, wherein the classifier directly converts the depth data into a clothing article size.

14. The method of claim 9, comprising updating a user profile with the clothing attributes, the clothing attributes corresponding to a clothing item object in a data structure of the user profile.

15. The method of claim 14, comprising searching for additional clothing items using the clothing attributes and the user profile.

16. The method of claim 15, wherein searching for additional clothing items includes applying a restriction that the additional clothing items differ from the clothing article in at least one of color, pattern, style, or clothing type.

17. At least one non-transitory machine readable medium including instructions for computer vision assisted item search, the instructions, when executed by a machine, cause the machine to perform operations comprising:

obtaining a composite image from a sensor array, the composite image including visual data and depth data;

filtering the composite image to isolate a clothing article represented in the composite image and determining a clothing type of the clothing article based on the depth data;

applying a classifier to the depth data to produce a set of clothing attributes that correspond to the clothing article based on the depth data and clothing type; and

providing the set of clothing attributes to a remote device.

18. The non-transitory machine readable medium of claim 17, wherein the set of clothing attributes include at least one of a size, a shape, or a texture.

19. The non-transitory machine readable medium of claim 18, wherein applying the classifier to the depth data to produce the set of clothing attributes includes:

determining a texture of a fabric using the depth data; and

identifying a material of the fabric that is correlated to the texture, and wherein the texture and material are included in the set of clothing attributes.

20. The non-transitory machine readable medium of claim 18, wherein the classifier scales a representation of the clothing article in the composite image using the depth data to produce an estimate of absolute measurements of the clothing article, and wherein the classifier converts the estimate of absolute measurements to a clothing article size.

21. The non-transitory machine readable medium of claim 18, wherein the classifier directly converts the depth data into a clothing article size.

22. The non-transitory machine readable medium of claim 17, wherein the operations comprise updating a user profile with the clothing attributes, the clothing attributes corresponding to a clothing item object in a data structure of the user profile.

23. The non-transitory machine readable medium of claim 22, wherein the operations comprise comprising searching for additional clothing items using the clothing attributes and the user profile.

24. The non-transitory machine readable medium of claim 23, wherein searching for additional clothing items includes applying a restriction that the additional clothing items differ from the clothing article in at least one of color, pattern, style, or clothing type.