Automated Character Model Generation via Neural Feature Extraction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Generating a customized 3D character model in computer graphics applications is difficult and time-consuming, requiring artistic competence and manual manipulation of high-dimensional parameters, making it challenging for users to create an accurate representation of a desired character.

Innovation Solution

A system and method using neural network models to process input images of a reference subject, identifying features and generating parameter vectors to create a parameterized character model, allowing for automated extraction and combination of relevant parameters to construct a character model matching the reference image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If manual manipulation of 3D model parameters is used to create a customized character, then the character can be customized according to user preference, but the process becomes time-consuming and difficult requiring artistic competence

Engineering Contradiction:

Improvecharacter customization capabilityVSAvoidtime to generate character model

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent replaces the manual mechanical manipulation of 3D model parameters with an automated image processing system using neural networks. The system automatically extracts character features from input images and generates corresponding 3D model parameters without requiring manual artistic manipulation, thereby reducing time consumption while maintaining customization capability

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables the character model generation process to serve itself by automatically processing input images through neural networks to extract features and generate parameters. The automated pipeline eliminates the need for human artistic intervention, allowing users to obtain customized character models quickly by simply providing reference images

Inventive Principle:

Principle #25Self-service

2Adaptability or versatility

If manual manipulation of 3D model parameters is used to create a customized character, then the character can be customized according to user preference, but artistic competence is required to obtain good results

Engineering Contradiction:

Improvecharacter customization capabilityVSAvoidease of character creation

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent replaces manual artistic manipulation with an automated neural network-based image processing system. The system automatically performs feature extraction, parameter generation, and model construction, eliminating the need for users to possess artistic competence while maintaining full customization capability through image-based input

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If intuitive direct on-screen manipulation of parameters is used, then user control is maintained, but fine tuning the appearance to match a reference is difficult due to high-dimensional parameter space

Engineering Contradiction:

Improveuser control over character parametersVSAvoidaccuracy of character appearance matching

Core Design Contradiction:

Ease of operationVSManufacturing precision

Solution Approach 1:

The patent replaces manual parameter manipulation with an automated neural network system that directly maps image features to 3D model parameters. This substitution enables precise matching of character appearance to reference images by leveraging the neural network's ability to navigate high-dimensional parameter spaces, while users retain control through image selection and system feedback

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The neural network acts as an intermediary between the input reference image and the 3D character model parameters. It automatically translates visual features from the image into corresponding model parameters, bridging the gap between user intent (represented by the reference image) and the complex high-dimensional parameter space of the 3D model, thereby achieving accurate appearance matching

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11648477B2Systems and methods for generating a model of a character from one or more images

Publication Date: 2023.05.16 ELECTRONIC ARTS INC

AI summary

A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.