Avatar Generation Using Generative Neural Networks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing messaging systems face challenges in efficiently generating and matching personalized avatars with text messages, requiring users to spend time creating customized media content.

Innovation Solution

The system uses generative neural networks (GNNs) trained with generative adversarial networks (GANs) to generate personalized avatars based on user input, including text placement characteristics such as font type, size, and mood, and matches these avatars with text messages based on determined characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If users manually create customized media content items, then personalization is achieved, but time consumption increases

Engineering Contradiction:

ImprovepersonalizationVSAvoidtime consumption

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system enables self-service by automatically generating customized media content items using the user's existing avatar and the text of their message. The generator creates personalized stickers without user intervention, allowing users to receive customized content instantly rather than manually creating it.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Users create their personalized avatar in advance, which is then stored and reused for generating multiple customized media content items. This preliminary action eliminates the need to recreate personalization elements for each new message, significantly reducing time consumption while maintaining personalization.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If the system generates customized avatars rapidly, then productivity increases, but complexity of the generation system increases

Engineering Contradiction:

Improveavatar generation speedVSAvoidgeneration system complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system uses copying by taking the user's existing avatar image and generating new customized versions through digital manipulation and composition. Rather than creating avatars from scratch, the system copies and transforms existing avatar elements, which simplifies the generation process while maintaining high productivity.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The avatar generator is designed as a universal system that can create multiple types of customized media content (stickers, emojis, animated avatars) from a single user avatar. This multi-functionality consolidates what would otherwise require multiple separate tools, reducing overall system complexity while enabling rapid generation of diverse personalized content.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250069306A1Customizable avatar generation system

Publication Date: 2025.02.27 SNAP INC

AI summary

Systems, methods, and computer readable media for customizable avatar generation system, where the methods include accessing text data, processing, using at least one processor, the text data to determine first characteristics of the text data, selecting a personalized avatar of a plurality of personalized avatars for the text data based on matching the first characteristics with second characteristics of the plurality of personalized avatars, generating a customized avatar based on the text data and the selected personalized avatar, and causing the customized avatar to be displayed on a display of a computing device.