System and method for clustering products by combining attribute data with image recognition

Patent metadata

Field	Value
Publication number	US-2020034668-A1
Application number	US-201916521750-A
Country	US
Kind code	A1
Filing date	Jul 25, 2019
Priority date	Jul 26, 2018
Publication date	Jan 30, 2020
Grant date	—

Abstract

Systems, methods, and computer-readable storage media for categorizing items based on attributes of the item and a shape of the item, where the shape of the item is determined from an image of the item. An exemplary system configured as disclosed herein can receive a request to categorize an item, the item having a plurality of attributes, and receive an image of the item. The system can identify, via a processor configured to perform image processing, a shape of the item based on the image, and transform the plurality of attributes and the shape of the item, into a plurality of quantifiable values. The system can then categorize the item based on the quantifiable values.

First claim

We claim:

1 . A method comprising:
receiving a request to categorize an item, the item having a plurality of attributes; executing, via a processor, a first random forest classifier on the plurality of attributes, resulting in a first random forest model; executing, via the processor, a second random forest classifier on an image of the item, resulting in a second random forest model; combining the executed random forest model and the second random forest model into a Bayesian Model Combination (BMC); executing, via the processor, the BMC, to yield BMC results; and classifying, via the processor, the item as belonging to a classification based on the BMC results.

2 . The method of claim 1 , wherein execution of the BMC comprises creating all possible linear combinations of the first random forest model and the second random forest model.

3 . The method of claim 1 , wherein the first random forest model and the second random forest model have distinct classification results.

4 . The method of claim 1 , further comprising:
generating, via the processor, a list of items similar to the item based on the classification.

5 . The method of claim 1 , further comprising:
generating, via the processor, a convolutional neural network model of the item using a convolutional neural network.

6 . The method of claim 5 , further comprising:
creating a convex linear combination of the first random forest model, the second random forest model, and the convolutional neural network model, the convex linear combination having weights for each model proportional to an accuracy of each respective model, wherein the classifying of the item as belonging to the classification further comprises using the convex linear combination.

7 . The method of claim 1 , wherein the classifying of the item utilizes a weighted equation to generate a score based on the BMC results, wherein each quantifiable value in the quantifiable values has an associated weight, and wherein the classification for the item is determined based on the score.

8 . A system comprising:
a processor; and a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising:
receiving a request to categorize an item, the item having a plurality of attributes;
executing a first random forest classifier on the plurality of attributes, resulting in a first random forest model;
executing a second random forest classifier on an image of the item, resulting in a second random forest model;
combining the executed random forest model and the second random forest model into a Bayesian Model Combination (BMC);
executing the BMC, to yield BMC results; and
classifying the item as belonging to a classification based on the BMC results.

9 . The system of claim 8 , wherein execution of the BMC comprises creating all possible linear combinations of the first random forest model and the second random forest model.

10 . The system of claim 8 , wherein the first random forest model and the second random forest model have distinct classification results.

11 . The system of claim 8 , the computer-readable storage medium having additional instructions which, when executed by the processor, cause the processor to perform operations comprising:
generating a list of items similar to the item based on the classification.

12 . The system of claim 1 , the computer-readable storage medium having additional instructions which, when executed by the processor, cause the processor to perform operations comprising:
generating a convolutional neural network model of the item using a convolutional neural network.

13 . The system of claim 5 , the computer-readable storage medium having additional instructions which, when executed by the processor, cause the processor to perform operations comprising:
creating a convex linear combination of the first random forest model, the second random forest model, and the convolutional neural network model, the convex linear combination having weights for each model proportional to an accuracy of each respective model, wherein the classifying of the item as belonging to the classification further comprises using the convex linear combination.

14 . The system of claim 1 , wherein the classifying of the item utilizes a weighted equation to generate a score based on the BMC results, wherein each quantifiable value in the quantifiable values has an associated weight, and wherein the classification for the item is determined based on the score.

15 . A non-transitory computer-readable storage medium having instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:
receiving a request to categorize an item, the item having a plurality of attributes; executing a first random forest classifier on the plurality of attributes, resulting in a first random forest model; executing a second random forest classifier on an image of the item, resulting in a second random forest model; combining the executed random forest model and the second random forest model into a Bayesian Model Combination (BMC); executing the BMC, to yield BMC results; and classifying the item as belonging to a classification based on the BMC results.

16 . The non-transitory computer-readable storage medium of claim 15 , wherein execution of the BMC comprises creating all possible linear combinations of the first random forest model and the second random forest model.

17 . The non-transitory computer-readable storage medium of claim 15 , wherein the first random forest model and the second random forest model have distinct classification results.

18 . The non-transitory computer-readable storage medium of claim 15 , having additional instructions which, when executed by the computing device, cause the computing device to perform operations comprising:
generating a list of items similar to the item based on the classification.

19 . The non-transitory computer-readable storage medium of claim 15 , having additional instructions which, when executed by the computing device, cause the computing device to perform operations comprising:
generating a convolutional neural network model of the item using a convolutional neural network.

20 . The non-transitory computer-readable storage medium of claim 19 , having additional instructions which, when executed by the computing device, cause the computing device to perform operations comprising:
creating a convex linear combination of the first random forest model, the second random forest model, and the convolutional neural network model, the convex linear combination having weights for each model proportional to an accuracy of each respective model, wherein the classifying of the item as belonging to the classification further comprises using the convex linear combination.

Assignees

• Walmart Apollo Llc

Inventors

• Dutta Tulika
• Batham Mohit
• Das Amlan J

Classifications

• G06N20/20 Primary

  Ensemble learning · CPC title

• G06V10/7625

  Hierarchical techniques, i.e. dividing or merging patterns to obtain a tree-like representation; Dendograms · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06F18/24155

  Bayesian classification · CPC title