Parametric Eye Modeling and Iris Texturing for Realistic 3D Heads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating 3D character models in video games face challenges in accurately depicting realistic human eyes, particularly in scenarios where the topology and blendshapes differ, leading to unrealistic eye placement, shape, and features, which can disrupt immersion and engagement.
Innovation Solution
A system and method for generating a 3D model of a head using a parametric eye model with spherical coordinates, normalizing eye positions, and generating eye patch areas through linear equations, combined with differentiable rendering and iris texture modeling to account for light refraction, allowing for realistic eye representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual parameter updating is used to adapt character models between different topologies and blendshapes, then the character appearance can be adjusted to match the reference character, but the process becomes difficult and time-consuming, requiring artistic competence
Solution Approach 1:
The patent uses automated parameter estimation to copy the appearance characteristics of a reference character into a target character model with different topology. The system extracts appearance parameters from the reference character and automatically maps them to the target model's blendshapes, eliminating manual adjustment and reducing the need for artistic competence while maintaining adaptability across different topologies
Solution Approach 2:
The patent changes the approach from manual parameter adjustment to automated parameter estimation. By using image processing and machine learning to automatically determine the appearance parameters of the reference character and then mapping these parameters to the target model, the system achieves both adaptability and time efficiency without requiring artistic skill
2Adaptability or versatility
If manual parameter updating is used to adapt character models, then the character appearance can be matched to the reference, but artistic competence is required which increases the complexity of the process
Solution Approach 1:
The patent replaces the mechanical process of manual parameter adjustment with an automated computational system. Instead of requiring artists to manually tweak parameters, the system uses image processing algorithms, machine learning models, and automated parameter mapping to achieve the same result, thereby reducing process complexity while maintaining appearance matching capability
Solution Approach 2:
The system performs self-service by automatically extracting appearance parameters from reference images and independently mapping them to the target model without human intervention. The automated parameter estimation and mapping processes eliminate the need for artistic competence, making the system self-sufficient in achieving character appearance matching
3Manufacturing precision
If new blendshapes are created to fill gaps in representation, then the character appearance can be improved, but the process is time-consuming and may not completely fill the gap due to topology limitations
Solution Approach 1:
The patent applies preliminary action by pre-establishing a comprehensive set of appearance parameters and their corresponding blendshape mappings before character creation. The system pre-processes reference images to extract appearance characteristics and pre-maps these to the target model's blendshapes, so that when a character needs to be created, the appropriate blendshapes are already prepared and can be instantly applied without time-consuming creation processes
4Adaptability or versatility
If eyes are not normalized to a fixed distance apart, then the head model can represent different interpupillary distances, but the eyes may be placed incorrectly or act in nonrealistic manners breaking immersion
Solution Approach 1:
The patent uses parameter changes by normalizing eye position to a fixed interpupillary distance as a reference parameter. This normalization ensures realistic eye placement and behavior while maintaining the ability to represent different actual interpupillary distances through scaling transformations. The fixed distance parameter serves as a reliable baseline that prevents unrealistic eye movements and placements while preserving model flexibility through parameter scaling
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the creation of realistic and immersive 3D models of human eyes by automating the process of eye positioning, shape, and texture generation, reducing manual effort and ensuring accurate rendering of eye features.
Implementation Method 1
The eyes are normalized to be spaced a fixed distance apart from one another in the 3D model
Implementation Method 2
combined with differentiable rendering and iris texture modeling to account for light refraction
Data Source
AI summary
A method for generating a three-dimensional (3D) model of a head is disclosed. One or more images of the head are obtained and the head includes eyes. A parametric model for the eyes that includes a set of parameters is retrieved. Values are assigned for each parameter in the set of parameters of the parametric model for the eyes based on the one or more images. Eye patch areas of areas surrounding the eyes are generated based on the values of the parameters in the set of parameters of the parametric model for the eyes. The 3D model of the head that includes the eyes and the eye patch areas is generated. The eyes are normalized to be spaced a fixed distance apart from one another in the 3D model, and a size of the head in the 3D model is scaled based on the fixed distance between the eyes.