Constructing, evaluating, and improving a search string for retrieving images indicating item use

Patent metadata

Field	Value
Publication number	US-11645329-B2
Application number	US-201715856505-A
Country	US
Kind code	B2
Filing date	Dec 28, 2017
Priority date	Dec 28, 2017
Publication date	May 9, 2023
Grant date	May 9, 2023

Abstract

Examples of techniques for constructing, evaluating, and improving a search string for retrieving images are disclosed. In one example implementation according to aspects of the present disclosure, a computer-implemented method includes receiving, by a processing device, a plurality of images as search results returned based at least in part on a search string for an item in the form of a tuple including an item class, an action and an actor. The method further includes determining, by the processing device, whether the search string is effective at indicating a common item use based on image similarity. The method further includes, based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string.

First claim

What is claimed is:

1. A computer-implemented method for improving a search string, the method comprising:
receiving, by a processing device, a plurality of images as search results returned based at least in part on a search string for an item in the form of a tuple comprising an item class, an action and an actor, wherein the item class is identified from an ontology listing by identifying a placement of the item in the ontology listing by item name, wherein the item class is a class of items that precedes the item name in the ontology listing;
determining, by the processing device, whether the search string is effective at indicating an item use based on image similarity; and
based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string,
wherein the alternative search string is in a form of a tuple comprising an item class, an action, and an actor,
wherein generating the alternative search string further comprises using ontology siblings to generate the alternative search string, and
wherein using ontology siblings to generate the alternative search string further comprises using ontology siblings of the item class to choose the action and the actor associated with the item class.

2. The computer-implemented method of claim 1 , wherein determining whether the search string is effective at indicating the item use comprises:
calculating, by the processing device, a similarity score by comparing at least one of the plurality of images to at least one other of the plurality of images;
determining, by the processing device, whether the similarity score exceeds a threshold; and
based at least in part on determining that the similarity score does not exceed the threshold, indicating that the search string is ineffective at indicating the item use.

3. The computer-implemented method of claim 1 , wherein generating the alternative search string further comprises identifying that the item class has multiple common uses.

4. The computer-implemented method of claim 3 , wherein the item is identified as having multiple common uses when a class for the item is found in multiple places in the ontology listing.

5. The computer-implemented method of claim 3 , wherein the item class is identified as having multiple common uses when multiple actions or actors are identified for the item class based at least in part on using a n-gram technique.

6. The computer-implemented method of claim 3 , wherein the item class is identified as having multiple common uses when multiple actions or actors are identified for the item class based at least in part on using a word embeddings technique.

7. A system for improving a search string, the system comprising:
a memory comprising computer readable instructions; and
a processing device for executing the computer readable instructions for performing a method, the method comprising:
receiving, by the processing device, a plurality of images as search results returned based at least in part on a search string for an item in the form of a tuple comprising an item class, an action and an actor, wherein the item class is identified from an ontology listing by identifying a placement of the item in the ontology listing by item name, wherein the item class is a class of items that precedes the item name in the ontology listing;
determining, by the processing device, whether the search string is effective at indicating an item use based on image similarity; and
based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string,
wherein the alternative search string is in a form of a tuple comprising an item class, an action, and an actor,
wherein generating the alternative search string further comprises using ontology siblings to generate the alternative search string, and
wherein using ontology siblings to generate the alternative search string further comprises using ontology siblings of the item class to choose the action and the actor associated with the item class.

8. The system of claim 7 , wherein determining whether the search string is effective at indicating the item use comprises:
calculating, by the processing device, a similarity score by comparing at least one of the plurality of images to at least one other of the plurality of images;
determining, by the processing device, whether the similarity score exceeds a threshold; and
based at least in part on determining that the similarity score does not exceed the threshold, indicating that the search string is ineffective at indicating the item use.

9. The system of claim 7 , wherein generating the alternative search string further comprises identifying that the item class has multiple common uses.

10. The system of claim 9 , wherein the item is identified as having multiple common uses when a class for the item is found in multiple places in the ontology listing.

11. The system of claim 9 , wherein the item class is identified as having multiple common uses when multiple actions or actors are identified for the item class based at least in part on using a n-gram technique.

12. A computer program product for improving a search string, the computer program product comprising:
a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing device to cause the processing device to perform a method comprising:
receiving, by the processing device, a plurality of images as search results returned based at least in part on a search string for an item in the form of a tuple comprising an item class, an action and an actor, wherein the item class is identified from an ontology listing by identifying a placement of the item in the ontology listing by item name, wherein the item class is a class of items that precedes the item name in the ontology listing;
determining, by the processing device, whether the search string is effective at indicating an item use based on image similarity; and
based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string,
wherein the alternative search string is in a form of a tuple comprising an item class, an action, and an actor,
wherein generating the alternative search string further comprises using ontology siblings to generate the alternative search string, and
wherein using ontology siblings to generate the alternative search string further comprises using ontology siblings of the item class to choose the action and the actor associated with the item class.

Assignees

• IBM

Inventors

• Gattiker Anne E
• Kashyap Sujatha
• Nguyen Minh Ngoc Binh
• Thomas Samuel
• Li Kaipeng
• Hubregtsen Thomas

Classifications

• G06F16/583 Primary

  using metadata automatically derived from the content · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/2425

  Iterative querying; Query formulation based on the results of a preceding query · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title