Image analysis enhanced related item decision

• US 10,592,782 B2
• Filed: 01/22/2018
• Issued: 03/17/2020
• Est. Priority Date: 01/22/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Deluca, Lisa Seacat, Greenberger, Jeremy A.
• Primary Examiner(s): Dulaney, Kathleen Y
• Application Number:
  US15/876,357
• Publication Number:
  US 20190228270A1
• Time in Patent Office: 785 Days
• Field of Search: 382224, 382305, 382100, 705 267, 705 271, 707723
• CPC Class Codes: G06F 16/5846 using extracted text
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/241 relating to the classification model, e.g. parametric or non-parametric approaches
• CPC Class Codes: G06F 18/285 Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: G06V 10/7788 the supervisor being a human, e.g. interactive learning with a human teacher
• CPC Class Codes: G06V 10/87 using selection of the recognition techniques, e.g. of a classifier in a multiple classifier system

Abstract

Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: obtaining from a user one or more data queries; identifying a product of interest in response to the one or more data query; examining a plurality of product records to determine a set of related products that are related to the product of interest, wherein the examining includes performing image analysis to extract one or more product topic classifier from product image data representing one or more product; and providing one or more output in response to the examining.

Claims

1. A method comprising:

obtaining from a user a data query, wherein the data query is a text based data query entered by the user;

identifying a product of interest in response to the text based data query obtained from the user;

examining a plurality of product records to determine a set of related products that are related to the product of interest, wherein the examining includes performing image analysis to extract one or more product topic classifier from product image data representing one or more product, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more product topic classifier of the product of interest; and

providing one or more output in response to the examining, wherein the method includes identifying a candidate related product in response to the text based data query obtained from the user, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract a first list of product topic classifiers associated the product of interest, and wherein the performing image analysis includes subjecting a product image of the candidate related product to image analysis to extract a second list of product topic classifiers associated to the candidate related product, wherein the method includes performing dissimilarity processing to determine a dissimilarity between the product of interest and the candidate related product, wherein the performing the dissimilarity processing includes discovering product topic classifiers of the first list that are not matched to a product topic classifier of the second list, and ascertaining product topic classifiers of the first list that are matched to a product topic classifier of the second list, and providing a dissimilarity score using the discovering and the ascertaining, wherein the examining a plurality of product records to determine a set of related products that are related to the product of interest comprises including the candidate related product in the set of related products based on the dissimilarity score being less than a threshold, and wherein the providing one or more output in response to the examining includes displaying the candidate related product to the user on a user interface of the user based on the dissimilarity score being less than the threshold.

2. The method of claim of 1, wherein the performing image analysis includes subjecting a product image of a second candidate related product for the product of interest to image analysis to extract one or more product topic classifier of the second candidate related product.

3. The method of claim of 1, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more first product topic classifier of the product of interest, and wherein the performing image analysis includes subjecting a product image of a second candidate related product for the product of interest to image analysis to extract one or more second product topic classifier of the second candidate related product, and wherein the examining includes comparing first data including the one or more first topic classifier to second data including the one or more second topic classifier, and the wherein the providing the one or more output in response to the examining includes excluding the second candidate related product from a determined set of related products based on the comparing indicating that a dissimilarity of the second candidate related product to the product of interest exceeds a threshold.

4. The method of claim of 1, wherein the identifying the product of interest in response to the one or more data query includes processing product records stored in a data repository, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more product topic classifier of the product of interest, and wherein the one or more output include updating the data repository so that a product of the data repository includes at least one of the one or more product topic classifier of the product of interest.

5. The method of claim 1, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more first product topic classifier of the product of interest, and wherein the performing image analysis includes subjecting a product image of a second candidate related product for the product of interest to image analysis to extract one or more second product topic classifier of a candidate related product, and wherein the examining includes producing a dissimilarity score using first data including the one or more first topic classifier and second data including the one or more second topic classifier, and wherein the providing the one or more output in response to the examining includes excluding the second candidate related product from a determined set of related products based on the dissimilarity score indicating that a dissimilarity of the second candidate related product to the product of interest exceeds a threshold.

6. The method of claim 1, wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing one or more output to perform a first action in response to the dissimilarity score having a first value, and wherein the providing one or more output includes providing one or more output to perform a second action in response to a second dissimilarity score having a second value.

7. The method of claim 1, wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing one or more output to perform a first action in response to the dissimilarity score having a first value, and wherein the providing one or more output includes providing one or more output to perform a second action in response to a second dissimilarity score having a second value, wherein the first action is displaying data of the candidate related product in a display area of a display featuring related products, wherein the second action is suppressing data of the candidate related product in a display area of the display featuring related products related to the product of interest.

8. The method of claim 1, wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing one or more output to perform a first action in response to the dissimilarity score having a first value, and wherein the providing one or more output includes providing one or more output to perform a second action in response to a second dissimilarity score having a second value, wherein the first action is presenting a user performing a product of interest search a menu option to suppress displaying data of the candidate related product in a display area of a display featuring related products, wherein the second action is presenting an administrator user who has added record data of the candidate related product into a data repository a menu option to add a product topic classifier extracted by image analysis of an image of the candidate related product into a data record for the candidate related product within the data repository.

9. The method of claim 1, wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing one or more output to perform a first action in response to the dissimilarity score having a first value, wherein the method includes monitoring performance of the dissimilarity score to decision mapping knowledgebase and updating parameter values of the dissimilarity score to decision mapping knowledgebase by machine learning based on the monitoring, wherein the monitoring includes monitoring of actions with respect to webpage content produced using the dissimilarity score to decision mapping knowledgebase.

10. The method of claim 1, wherein the method includes providing the plurality of product records using a displayed administrator user interface that has an image selection area to facilitate selection of one or more image of a product by an administrator user, and also includes a keyword area configured to facilitate entry of keyword text by the administrator user, wherein the administrator user interface is configured to permit the administrator user to activate an image analysis process wherein a selected one or more image of a product by the administrator user selected using the image selection area is subject to image analysis to extract one or more product classifiers of a product represented in the selected one or more image, and wherein responsively to activation of the image analysis process by the administrator user auto-populated text specifying the one or more extracted product classifiers is auto-populated into the keyword area in a form that permits editing of the auto-populated text by the administrator user.

11. A computer program product comprising:

a computer readable storage medium readable by one or more processing circuit and storing instructions for execution by one or more processor for performing a method comprising;

obtaining from a user a data query, wherein the data query is a text based data query entered by the user;

identifying a product of interest in response to the text based data query obtained from the user;

examining a plurality of product records to determine a set of related products that are related to the product of interest, wherein the examining includes performing image analysis to extract one or more product topic classifier from product image data representing one or more product, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more product topic classifier of the product of interest; and

providing one or more output in response to the examining, wherein the method includes identifying a candidate related product in response to the text based data query obtained from the user, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract a first list of product topic classifiers associated the product of interest, and wherein the performing image analysis includes subjecting a product image of the candidate related product to image analysis to extract a second list of product topic classifiers associated to the candidate related product, wherein the method includes performing dissimilarity processing to determine a dissimilarity between the product of interest and the candidate related product, wherein the performing the dissimilarity processing includes discovering product topic classifiers of the first list that are not matched to a product topic classifier of the second list, and ascertaining product topic classifiers of the first list that are matched to a product topic classifier of the second list, and providing a dissimilarity score using the discovering and the ascertaining, wherein the examining a plurality of product records to determine a set of related products that are related to the product of interest comprises including the candidate related product in the set of related products based on the dissimilarity score being less than a threshold, and wherein the providing one or more output in response to the examining includes displaying the candidate related product to the user on a user interface of the user based on the dissimilarity score being less than the threshold.

12. The computer program product of claim 11, wherein the performing image analysis includes subjecting a product image of a second candidate related product for the product of interest to image analysis to extract one or more product topic classifier of the second candidate related product.

13. The computer program product of claim 11, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more first product topic classifier of the product of interest, and wherein the performing image analysis includes subjecting a product image of a second candidate related product for the product of interest to image analysis to extract one or more second product topic classifier of the second candidate related product, and wherein the examining includes comparing first data including the one or more first topic classifier to second data including the one or more second topic classifier, and wherein the providing the one or more output in response to the examining includes excluding the candidate related product from a determined set of related products based on the comparing indicating that a dissimilarity of the candidate related product to the product of interest exceeds a threshold.

14. The computer program product of claim 11, wherein the plurality of records are stored in a data repository and wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing one or more output to perform a first action in response to the dissimilarity score having a first value, and wherein the providing one or more output includes providing one or more output to perform a second action in response to a second dissimilarity score having a second value, wherein the first action is presenting a user performing a product of interest search a menu option to suppress displaying data of the candidate related product in a display area of a display featuring related products, wherein the second action is presenting an administrator user who has added record data of the candidate related product into the data repository a menu option to add a product topic classifier extracted by image analysis of an image of the candidate related product into a data record for the candidate related product within the data repository, wherein the method includes providing the plurality of product records using a displayed administrator user interface that has an image selection area to facilitate selection of one or more image of a product by the administrator user, and also includes a keyword area configured to facilitate entry of keyword text by the administrator user, wherein the administrator user interface is configured to permit the administrator user to activate an image analysis process wherein a selected one or more image of a product by the administrator user selected using the image selection area is subject to image analysis to extract one or more product classifiers of a product represented in the selected one or more image, and wherein responsively to activation of the image analysis process by the administrator user auto-populated text specifying the one or more extracted product classifiers is auto-populated into the keyword area in a form that permits editing of the auto-populated text by the administrator user.

15. The computer program product of claim 11, wherein the plurality of records are stored in a data repository and wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing one or more output includes providing the one or more output to perform a first action in response to the dissimilarity score having a first value, and wherein the providing the one or more output includes providing one or more output to perform a second action in response to a second dissimilarity score having a second value, wherein the first action is presenting a user performing a product of interest search a menu option to suppress displaying data of the candidate related product in a display area of a display featuring related products, wherein the second action is presenting an administrator user who has added record data of the candidate related product into the data repository a menu option to add a product topic classifier extracted by image analysis of an image of the candidate related product into a data record for the candidate related product of within the data repository, wherein the method includes providing the plurality of product records using a displayed administrator user interface that has an image selection area to facilitate selection of one or more image of a product by the administrator user, and also includes a keyword area configured to facilitate entry of keyword text by the administrator user, wherein the administrator user interface is configured to permit the administrator user to activate an image analysis process wherein a selected one or more image of a product by the administrator user selected using the image selection area is subject to image analysis to extract one or more product classifiers of a product represented in the selected one or more image, and wherein responsively to activation of the image analysis process by the administrator user auto-populated text specifying the one or more extracted product classifiers is auto-populated into the keyword area in a form that permits editing of the auto-populated text by the administrator user, wherein the providing the one or more output in response to the examining includes using a dissimilarity score to decision mapping knowledgebase that maps product of interest to candidate related product dissimilarity scores to decisions associated to the product of interest to candidate related product dissimilarity scores, wherein the providing the one or more output includes providing the one or more output to perform a first action in response to the dissimilarity score having a first value, wherein the method includes monitoring performance of the dissimilarity score to decision mapping knowledgebase and updating parameter values of the dissimilarity score to decision mapping knowledgebase by machine learning based on the monitoring, wherein the monitoring includes monitoring of actions with respect to webpage content produced using the dissimilarity score to decision mapping knowledgebase.

16. A system comprising:

a memory;

at least one processor in communication with the memory; and

program instructions executable by one or more processor via the memory to perform a method comprising;

obtaining from user a data query, wherein the data query is a text based data query entered by the user;

identifying a product of interest in response to the text based data query obtained from the user;

examining a plurality of product records to determine a set of related products that are related to the product of interest, wherein the examining includes performing image analysis to extract one or more product topic classifier from product image data representing one or more product, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more product topic classifier of the product of interest; and

providing one or more output in response to the examining, wherein the method includes identifying a candidate related product in response to the text based data query obtained from the user, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract a first list of product topic classifiers associated the product of interest, and wherein the performing image analysis includes subjecting a product image of the candidate related product to image analysis to extract a second list of product topic classifiers associated to the candidate related product, wherein the method includes performing dissimilarity processing to determine a dissimilarity between the product of interest and the candidate related product, wherein the performing the dissimilarity processing includes discovering product topic classifiers of the first list that are not matched to a product topic classifier of the second list, and ascertaining product topic classifiers of the first list that are matched to a product topic classifier of the second list, and providing a dissimilarity score using the discovering and the ascertaining, wherein the examining a plurality of product records to determine a set of related products that are related to the product of interest comprises including the candidate related product in the set of related products based on the dissimilarity score being less than a threshold, and wherein the providing one or more output in response to the examining includes displaying the candidate related product to the user on a user interface of the user based on the dissimilarity score being less than the threshold.

17. The system of claim of 16, wherein the method includes identifying a second candidate related product in response to the text based data query obtained from the user, wherein the performing image analysis includes subjecting a product image of the product of interest to image analysis to extract one or more first product topic classifier of the product of interest, and wherein the performing image analysis includes subjecting a product image of the second candidate related product for the product of interest to image analysis to extract one or more second product topic classifier of the second candidate related product, and wherein the examining includes comparing first data including the one or more first topic classifier to second data including the one or more second topic classifier, and wherein the providing the one or more output in response to the examining includes excluding the candidate related product from a determined set of related products based on the comparing indicating that a dissimilarity of the candidate related product to the product of interest exceeds a threshold.