Item identification among item variations
This disclosure describes a system for automatically identifying an item from among a variation of items of a same type. For example, an image may be processed and resulting item image information compared with stored item image information to determine a type of item represented in the image. If the matching stored item image information is part of a cluster, the item image information may then be compared with distinctive features associated with stored item image information of the cluster to determine the variation of the item represented in the received image.
1. A computing system, comprising:
one or more processors; and
a memory coupled to the one or more processors and storing program instructions that when executed by the one or more processors cause the one or more processors to:
receive an image of an item located within a materials handling facility;
process the image of the item to generate a first item image information, wherein the first item image information includes:
a plurality of features of the item identified in the image;
an arrangement of the plurality of features; and
wherein the program instructions that when executed by the one or more processors to cause the one or more processors to process the image, further include program instructions that cause the one or more processors to:
determine a correlation score between the first item image information and a second item image information, wherein the second item image information is associated with the item and accessible from an item images data store;
determine that the correlation score exceeds a threshold;
determine that the second item image information is associated with a set, wherein the set includes the second item image information and a third item image information, and wherein the second item image information includes a distinctive feature identifying a difference between the second item image information and the third item image information;
determine a distinctive features correlation score between the first item image information and the distinctive feature;
determine that the distinctive features correlation score exceeds a distinctive features threshold; and
identify the item as the item associated with the second item image information.
2. The computing system of claim 1 , wherein the distinctive feature identifies a feature of the second item image information that is different than a feature of the third item image information.
3. The computing system of claim 1 , wherein the feature of the item is at least one of a shape of the item, a color of the item, a logo on the item, a label on the item, a word on the item, a symbol on the item, a character on the item, a number on the item, a texture of the item, a gradient of the item, a shape of the item, a reflectivity of the item, or an edge of the item.
4. The computing system of claim 1 , wherein the second item image information further includes at least one of the image, a time the image was captured, a location within the materials handling facility where the image was captured, a boundary of the item within the image, or an image capture device identifier identifying the image capture device that captured the image.
5. The computing system of claim 1 , wherein the arrangement of features of the item includes a two dimensional relationship of a feature of a plurality of features of the item with respect to other features of the plurality of features of the item.
6. The computing system of claim 1 , wherein the arrangement of features of the item includes a three-dimensional relationship of a feature of a plurality of features of the item with respect to other features of the plurality of features of the item.
7. The computing system of claim 1 , wherein at least one of the plurality of features of the item includes one or more of a size, a shape, a color, a wording, or a graphic.
8. A computer-implemented method for automatically updating an item images data store for use in identifying items, comprising:
under control of one or more computing systems configured with executable instructions, receiving an image of an item;
processing the image of the item to generate a first item image information that includes an arrangement of features of the item;
for each of a first plurality of stored item image information, comparing the arrangement of features with a stored arrangement of features associated with a stored item image information of the first plurality of stored item image information;
determining that a highest correlation score exceeds a threshold;
comparing the first item image information with a second item image information associated with a stored arrangement of features with the highest correlation score to determine a distinctive feature of the first item image information that is different than a feature of the second item image information;
associating the distinctive feature with the first item image information;
determining a distinctive feature correlation score between the first item image information and the second item image information;
determining that the distinctive feature correlation score exceeds a distinctive features threshold; and
defining a cluster that includes the first item image information and the second item image information.
9. The computer-implemented method of claim 8 , further comprising:
comparing the first item image information and the second item image information to identify a plurality of common features; and
identifying the distinctive feature as a feature of the first item image information that is not a common feature.
10. The computer-implemented method of claim 8 , wherein the arrangement of the identified plurality of features is a two-dimensional arrangement of identified plurality of features with respect to each other and with respect to the image.
11. The computer-implemented method of claim 8 , wherein:
identifying a plurality of features includes identifying a plurality of distinctive features; and
the arrangement includes an arrangement of the plurality of distinctive features.
12. The computer-implemented method of claim 8 , further comprising:
prior to comparing the first item image information with the second item image information to identify the distinctive feature:
determining a correlation score between the first item image information and the second item image information; and
determining that the correlation score exceeds a threshold.
13. The computer-implemented method of claim 8 , wherein the second item image information is stored in an item images data store and is associated with the second item.
14. The computer-implemented method of claim 8 , wherein comparing the first item image information with a second item image information further includes:
identifying a second distinctive feature of the second image that is different than a feature of the first image; and
associating the second distinctive feature with the second item image information.
15. The computer-implemented method of claim 14 , wherein the second distinctive feature identifies features of the second item image information that are unique to the second item image information when compared to the first item image information.
16. A computing system, comprising:
one or more processors; and
a memory coupled to the one or more processors and storing program instructions that when executed by the one or more processors cause the one or more processors to:
receive an image of an item;
process the image to determine an arrangement of features of the item;
for each of a first plurality of stored item image information:
compare the arrangement of features with a stored arrangement of features associated with a stored item image information of the first plurality of stored item image information;
determine a correlation score between the arrangement of features and the stored arrangement of features;
determine that a highest correlation score exceeds a threshold;
determine that a stored item image information of the first plurality of item image information associated with a stored arrangement of features with the highest correlation score is associated with a cluster, wherein the cluster identifies a second plurality of stored item image information;
for each of the second plurality of stored item image information:
compare the item image information with a distinctive feature associated with the stored item image information of the second plurality of stored item image information;
determine a distinctive features correlation score; and
identify the item in the image as an item associated with the stored item image information of the second plurality of stored item image information associated with a distinctive feature having a highest distinctive features correlation score.
17. The computing system of claim 16 , wherein the image of the item is captured while the item is being removed from an inventory location or placed at the inventory location.
18. The computing system of claim 16 , wherein the item image information further includes an identification of an image capture device that obtained the image, an identification of a time when the image was captured, or information on pan-tilt-zoom information for the image capture device.
19. The computing system of claim 16 , wherein the stored item image information associated with the cluster represents a plurality of items of a same type.
20. The computing system of claim 19 , wherein the distinctive feature identifies visual differences between the items.
21. The computing system of claim 16 , wherein the distinctive feature identifies an arrangement of features associated with a stored item image information associated with the second plurality of item image information that is different than an arrangement of features associated with at least one other stored item image information associated with the second plurality of item image information.
22. The computing system of claim 16 , wherein the second plurality of stored item image information is a subset of the first plurality of stored item image information.
23. A computing system, comprising:
one or more processors; and
a memory coupled to the one or more processors and storing program instructions that when executed by the one or more processors cause the one or more processors to at least:
receive an image of an item;
process the image to determine an arrangement of features of the item;
compare the arrangement of features with a stored arrangement of features associated with a stored item image information;
determine a correlation score between the arrangement of features and the stored arrangement of features;
determine that the correlation score exceeds a threshold;
determine that the stored item image information is associated with a second stored item image information;
compare the item image information with a distinctive feature associated with the second stored item image information;
determine a distinctive features correlation score; and
identify, based at least in part on the distinctive features correlation score, the item in the image as an item associated with at least one of the stored item image information or the second stored item image information.
24. The computing system of claim 23 , wherein the program instructions that when executed by the one or more processors further cause the one or more processors to at least:
associate the features of the item with at least one of the stored item image information or the second stored item image information.