Text Classification for Graphical Element Selection in Digital Media Templates

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Designing digital media content, such as web pages, often requires expertise that is inaccessible to many users, as they struggle to select appropriate graphical elements that match the context, emotions, and entities of the text being presented.

Innovation Solution

A computer program and method that uses classifications from text to determine instances of graphical element types through a machine learning module, recommending styling and appearance based on design expertise, and allowing users to select these elements for optimal presentation in a template layout.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If users design digital media content using traditional tools and templates, then they have access to basic layout capabilities, but they lack the expertise to select appropriate graphical elements that match text context, emotions, and entities

Engineering Contradiction:

Improveease of selecting appropriate graphical elementsVSAvoidprecision of matching graphical elements to text context

Core Design Contradiction:

Ease of operationVSManufacturing precision

Solution Approach 1:

The system performs automatic classification of text into context, emotions, and entities without requiring user expertise. The machine learning module autonomously selects recommended graphical elements based on the classified text, enabling the system to serve itself by generating design recommendations without human design expertise.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual design expertise (mechanical human judgment) with an automated machine learning system. The classification module and recommendation engine substitute for human designers' ability to match graphical elements with text context, emotions, and entities, providing expert-level precision through computational algorithms rather than human intuition.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Manufacturing precision

If professional designers are used to create digital media content, then high-quality design with appropriate graphical elements is achieved, but design expertise becomes inaccessible to many users

Engineering Contradiction:

Improvequality of design element selectionVSAvoidaccessibility of design expertise to users

Core Design Contradiction:

Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The machine learning recommendation system acts as an intermediary between the user and professional design expertise. Instead of requiring direct access to human designers, the system mediates by automatically analyzing text and generating design recommendations that embody professional design principles, making expert-level quality accessible to ordinary users through an automated intermediary layer.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system captures and replicates professional design expertise in the form of machine learning models trained on design principles and patterns. By copying the knowledge and judgment of professional designers into an automated system, the patent enables widespread access to expert-level design quality without requiring users to have direct access to human experts.

Inventive Principle:

Principle #26Copying

3Adaptability or versatility

If automated systems are used to generate design recommendations, then design expertise becomes accessible to all users, but the system requires complex machine learning modules and text classification capabilities

Engineering Contradiction:

Improveaccessibility of design tools to usersVSAvoidcomplexity of machine learning and classification systems

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system divides the complex design assistance task into separate functional modules: a text classification module that analyzes text for context, emotions, and entities, and a recommendation module that generates graphical element suggestions based on classifications. This segmentation allows each module to specialize in specific functions, managing overall system complexity through modular architecture while providing comprehensive design assistance.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11200366B2Using classifications from text to determine instances of graphical element types to include in a template layout for digital media output

Publication Date: 2021.12.14 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Provided are a computer program product, system, and method for using classifications from text to determine instances of graphical element types to include in a template layout for digital media output. Text is processed to determine classifications. The determined classifications of the text are inputted to a machine learning module to output instances for graphical element types. The outputted instances of the graphical element types are rendered in a user interface for a user to select. User selection is received of one of the instances rendered in the user interface for each of the graphical element types. The text with the user selected instances for the graphical element types are rendered in a template layout. The template layout including the text rendered with the user selected instances for the graphical element types is rendered in an output layout of digital media.