Parts-Based Muscle Deformation for Novel Pose Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for creating realistic digital human models in the entertainment industry are highly manual, time-consuming, and expensive, as they require hand-crafting Computer Graphics (CG) artwork or using 3D/4D scanners that fail to capture novel face expressions and body actions efficiently.
Innovation Solution
A muscle deformation method using a parts-based approach that acquires and tracks 4D scanned Range of Motion (ROM) shapes, performing deformation transfer specific to poses based on vertex correspondences, allowing for the generation of novel poses and reducing manual processing by segmenting and tracking muscle groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If hand-crafting CG artwork or using conventional 3D/4D scanners is used to create digital human models, then realistic models can be created, but the process is highly manual, time-consuming and expensive
Solution Approach 1:
The patent segments the human body into multiple muscle groups (e.g., biceps, triceps, quadriceps, hamstrings) and processes each muscle group independently. This segmentation allows the system to capture and deform specific muscle regions separately, enabling efficient generation of realistic muscle deformations without processing the entire body as one unit, thus improving both quality and efficiency
Solution Approach 2:
The patent uses 4D scanning to capture real muscle deformation data from actual human subjects and creates digital copies of these deformations. These captured deformation patterns are then reused and applied to generate realistic muscle movements in virtual characters, eliminating the need for manual crafting while maintaining high realism
2Reliability
If 4D scanner studios capture natural surface dynamics with fixed videos and actions, then surface movement is captured, but novel face expressions or body actions cannot be created
Solution Approach 1:
By segmenting the body into independent muscle groups with distinct vertex correspondences, the system can selectively activate and deform specific muscle regions to generate novel poses and expressions that were not captured in the original scanning session, while maintaining reliable surface dynamics in captured regions
Solution Approach 2:
The patent implements a dynamic deformation transfer system that can adaptively apply muscle deformations to generate novel poses. The system uses pose-specific deformation transfer based on vertex correspondences to create realistic muscle movements for actions and expressions beyond those originally captured, enhancing versatility while preserving capture reliability
3Quantity of substance
If dummy actors perform many sequences of actions for scanning, then comprehensive motion data is captured, but the workload is huge
Solution Approach 1:
The system captures muscle deformation patterns from limited actual performances and creates reusable digital copies of these deformation patterns. These copied deformation data can then be applied to generate comprehensive motion data for multiple poses and actions without requiring the actor to physically perform every sequence, dramatically reducing workload while maintaining data completeness
Solution Approach 2:
The patent uses parameter-based deformation transfer that allows the same captured muscle deformation data to be adapted and applied to various poses and actions by changing transformation parameters. This enables comprehensive motion data generation from limited capture sessions, reducing the time and effort required for extensive actor performances
Data Source
AI summary
Parts-based blendshape generation involves establishing a 4D capture sequence. A face, hands and legs are able to be established for the meshes using a template to generate tracked and templated meshes. Specific muscle deformations are extracted. Pose-specific spatial deformations are integrated into a pose to generate specific flesh deformations. A muscle deformation method enables muscle part-based approach deformation (e.g., just biceps if flexing arms). The muscle deformation also enables generating a novel pose not captured.