Image context based text generation
AssigneeShutterstock, Inc.
InventorsRaymond Lúí Smyth, Christopher Loy
Life of this patent
Estimating the term…
Abstract
Methods, systems, and storage media for generating contextually relevant text from image descriptions and user intent are disclosed. Exemplary implementations may: receive an image and a user-defined intent for text output; analyze the received image to generate a contextual description of the image; generate a query based on the contextual description of the image and the user-defined intent; and generate the text output based on the query.
Claims20
A method for generating contextually relevant text from image descriptions and user intent, comprising:
- receiving an image and a user-defined intent for text output, the user-defined intent including at least a description of an intended use, a selected tone, and a selected content type for the text output;
- generating a system message including at least a first text argument based on the user-defined intent for the text output;
- analyzing the received image to generate a contextual description of the image;
- modifying, in response to a selection for an image influence feature, the system message by prepending a second text argument including the contextual description of the image to the first text argument, wherein the image influence feature enables consideration of the image;
- generating a query based on the system message;
- and
- generating the text output based on the query, wherein the text output is formatted and structured according to a set of guidelines corresponding to the selected content type, and contents of the text output align with a semantic, stylistic, and thematic characteristics of the image.
A system configured for generating contextually relevant text from image descriptions and user intent, the system comprising:
- one or more hardware processors configured by machine-readable instructions to:
- receive an image and a user-defined intent for text output, the user-defined intent including at least a description of an intended use, a selected tone, and a selected content type for the text output;
- generate a system message including at least a first text argument based on the user-defined intent for the text output;
- analyze the received image to generate a contextual description of the image;
- modify, in response to a selection for an image influence feature, the system message by prepending a second text argument including the contextual description of the image to the first text argument, wherein the image influence feature enables consideration of the image;
- generate a query based on the system message;
- and
- generate the text output based on the query, wherein the text output is formatted and structured according to a set of guidelines corresponding to the selected content type, and contents of the text output align with a semantic, stylistic, and thematic characteristics of the image.
A non-transient computer-readable storage medium having instructions embodied thereon, the instructions being executable by one or more processors to perform a method for generating contextually relevant text from image descriptions and user intent, the method comprising:
- receiving an image and a user-defined intent for text output, the user-defined intent including at least a description of an intended use, a selected tone, and a selected content type for the text output;
- generating a system message including at least a first text argument based on the user-defined intent for the text output;
- analyzing the received image to generate a contextual description of the image;
- modifying, in response to a selection for an image influence feature, the system message by prepending a second text argument including the contextual description to the first text argument, wherein the image influence feature enables consideration of the image;
- generating a query based on the system message;
- and
- generating the text output based on the query, wherein the text output is formatted and structured according to a set of guidelines corresponding to the selected content type, and contents of the text output align with a semantic, stylistic, and thematic characteristics of the image.
Dependencies are inferred from the claim text.