Systems and methods for determining similarity of online items

Patent metadata

Field	Value
Publication number	US-12518308-B2
Application number	US-202217820401-A
Country	US
Kind code	B2
Filing date	Aug 17, 2022
Priority date	Aug 17, 2022
Publication date	Jan 6, 2026
Grant date	Jan 6, 2026

Abstract

Disclosed are methods and systems for determining similarity of online items. For instance, an identifier of a first web page associated with a first item navigated to by a browser may be received from a browser extension application executing on a computing device. Image(s) and text may be extracted from the first web page and provided as input to a machine learning model along with image(s) and text extracted from a second web page of a plurality of web pages associated with a second item of a plurality of items that are stored in a data store. A probability at/above a predefined threshold that the first and second items are the same item may be received as output. A notification indicating identification of the second item as the same item and including information associated with the second item may be generated and provided to the browser extension application for display.

First claim

What is claimed is:

1 . A computer-implemented method for determining similarity of online items, comprising:
extracting and storing, in a data store, one or more images and text from each web page of a plurality of web pages associated with an item of a plurality of items, wherein the one or more images and text from each web page are stored relative to one another according to a data structure; receiving, from a browser extension application executing on a computing device, an identifier of a first web page navigated to by a browser executing on the computing device, wherein the first web page is associated with a first item; extracting one or more images and text from the first web page; receiving, from the data store, one or more images and text from a second web page of the plurality of web pages associated with a second item of the plurality of items; providing the one or more images and the text from the first web page and the one or more images and the text from the second web page as input to a machine learning model trained to predict whether the first item and the second item are a same item; receiving, as output from the machine learning model, a probability at or above a predefined threshold that the first item and the second item are the same item; generating a notification indicating identification of the second item as the same item as the first item and including information associated with the second item; selecting, from the plurality of web pages, a third web page associated with a third item from the plurality of items for which one or more images and text of the third web page are provided as input to the machine learning model along with the one or more images and the text from the first web page to predict whether the first item and the third item are a same item, wherein:
the selection is based on a position of the third web page relative to the second web page within the data structure and the prediction that the first item and the second item are the same item, and
the notification is updated to indicate identification of the third item as the same item as the first item and including information associated with the third item if when the first item and the third item are predicted to be the same item; and
providing the notification to the browser extension application for display on the computing device.

2 . The method of claim 1 , wherein providing the one or more images and the text from the first web page and the one or more images and the text from the second web page as input to the machine learning model comprises:
generating a plurality of vectors, including a vector for each of the one or more images from the first web page, a vector for the text from the first web page, a vector for each of the one or more images from the second web page, and a vector for the text from the second web page; and providing the plurality of vectors as the input to the machine learning model.

3 . The method of claim 1 , wherein the extracting and storing of the one or more images and the text from each web page of the plurality of web pages comprises performing web scraping.

4 . The method of claim 1 , further comprising:
receiving, as further output from the machine learning model, a confidence level associated with the probability, wherein the probability and the confidence level are at or above the predefined threshold.

5 . The method of claim 1 , wherein at least a portion of the information in the notification includes or is based on at least one of the one or more images from the second web page or the text from the second web page.

6 . The method of claim 1 , further comprising:
receiving consumer data associated with an account of the browser extension application, wherein at least a portion of the information associated with the second item in the notification is customized based on the consumer data.

7 . The method of claim 6 , wherein the consumer data includes one or more of consumer location, consumer preference data, or consumer interaction data.

8 . The method of claim 1 , wherein the information associated with the second item included in the notification comprises at least one of: the probability that the first item and the second item are the same item, a confidence level associated with the probability, a name of the second item, a manufacturer of the second item, an image of the second item, a description of the second item, a review of the second item, a price of the second item, an estimated shipping cost associated with the second item, an estimated time of delivery of the second item, a merchant associated with the second web page, an offer associated with the second item or the merchant, or a navigation link to the second web page.

9 . The method of claim 1 , wherein the data structure comprises a binary tree structure.

10 . The method of claim 1 , wherein when the first item and the third item are predicted to be the same item, the notification is updated to further indicate identification of the third item as the same item as the first item and further include information associated with the third item.

11 . The method of claim 10 , further comprising:
ranking the second item and the third item based on the probability that the first item and the second item are the same item and the probability that the first item and the third item are the same item; and ordering the information associated with second item and the third item within the notification based on the ranking.

12 . The method of claim 11 , wherein the ranking is further based on consumer data associated with an account of the browser extension application, the consumer data including one or more of consumer location, consumer preference data, or consumer interaction data.

13 . A system, the system comprising:
at least one memory storing instructions; and at least one processor operatively connected to the at least one memory, and configured to execute the instructions to perform operations for determining similarity of online items, the operations including:
extracting and storing, in a data store, one or more images and text from each web page of a plurality of web pages associated with an item of a plurality of items, wherein the one or more images and text from each web page are stored relative to one another according to a data structure;
receiving, from a browser extension application executing on a computing device, an identifier of a first web page navigated to by a browser executing on the computing device, wherein the first web page is associated with a first item;
extracting one or more images and text from the first web page;
receiving, from the data store, one or more images and text from a second web page of the plurality of web pages associated with a second item of the plurality of items;
providing the one or more images and the text from the first web page and the one or more images and the text from the second web page as input to a machine learning model trained to predict whether the first item and the second item are a same item;
receiving, as output from the machine learning model, a probability at or above a predefined threshold that the first item and the second item are the same item;
generating a notification indicating identification of the second item as the same item as the first item and including information associated with the second item;
selecting, from the plurality of web pages, a third web page associated with a third item from the plurality of items for which one or more images and text of the third web page are provided as input to the machine learning model along

Assignees

• Capital One Services Llc

Inventors

• Yee Bryant
• Mcclanahan Brian
• Vargas Cruz

Classifications

• G06Q30/0629 Primary

  by pre-processing results, e.g. ranking or ordering results · CPC title