Generative Neural Network Animation Transfer System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital animation systems require manual user interaction to define how objects move and interact with their environment, leading to inefficiencies and increased computational costs, as they lack a mechanism for transferring animation between objects.
Innovation Solution
A computing device implements a system using generative neural networks, including a meshing module, warping module, and training module, which uses a generative adversarial network to automatically transfer animation from one object to another by mapping features and deforming meshes, eliminating the need for manual user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual user interaction is used to define animation for each object, then animation quality can be controlled, but the time and computational cost increase significantly
Solution Approach 1:
The system creates a source animation from reference images and copies it to multiple target objects. The animation definition is generated once and then transferred to numerous objects without requiring manual re-creation, dramatically reducing time while maintaining consistent quality across all animated objects
Solution Approach 2:
The system automatically generates animation definitions by processing reference images through neural networks. The animation creation process serves itself by using the generated animation as a template for subsequent objects, eliminating the need for continuous manual intervention and reducing both time and computational costs
2Manufacturing precision
If animation definitions are created for every object individually, then animation quality is maintained, but computational cost increases
Solution Approach 1:
A single animation definition generated from reference images serves multiple target objects universally. The same animation template can be applied to different objects (characters, animals, robots) without re-computation, reducing computational cost while maintaining quality through the universal animation definition
Solution Approach 2:
The system performs preliminary action by generating the animation definition once from reference images before applying it to multiple objects. This pre-computed animation definition is then reused across numerous targets, avoiding redundant computational operations and reducing overall energy consumption
3Adaptability or versatility
If conventional animation systems are used, then each object can be animated independently, but no mechanism exists to transfer animation between objects
Solution Approach 1:
The system introduces an intermediary animation definition that acts as a bridge between source reference images and target objects. This intermediate representation enables seamless transfer of animation characteristics across different objects while maintaining manageable system complexity through standardized processing pipelines
Data Source
AI summary
In implementations of object animation using generative neural networks, one or more computing devices of a system implement an animation system for reproducing animation of an object in a digital video. A mesh of the object is obtained from a first frame of the digital video and a second frame of the digital video having the object is selected. Features of the object from the second frame are mapped to vertices of the mesh, and the mesh is warped based on the mapping. The warped mesh is rendered as an image by a neural renderer and compared to the object from the second frame to train a neural network. The rendered image is then refined by a generator of a generative adversarial network which includes a discriminator. The discriminator trains the generator to reproduce the object from the second frame as the refined image.