Virtual Character Rendering Level of Detail Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in reducing the rendering cost of virtual characters in virtual reality environments, especially as the number of virtual characters increases, leading to high computational demands even for high-end devices.
Innovation Solution
A real-time rendering generating apparatus and method that determines rendering levels for virtual characters based on classification rules related to their importance and motion data, using character levels of detail corresponding to customized body parts and skeletal models, allowing for optimized rendering resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the number of virtual characters is increased, then the virtual reality world becomes more populated and interactive, but the rendering cost increases significantly
Solution Approach 1:
The patent segments virtual characters into multiple level of detail (LOD) versions, where each LOD represents a different level of geometric complexity. The system divides characters into head, body, and limb components that can be independently rendered at different LODs, allowing the renderer to selectively apply high-detail rendering only to visible and important character portions while using simplified models for others, thus reducing overall rendering cost while maintaining character population.
Solution Approach 2:
The patent applies local quality by rendering different body parts of virtual characters at different levels of detail based on their visual importance and visibility. Critical regions such as the head and upper body are rendered with higher geometric fidelity, while less important or occluded regions use simplified meshes. This selective rendering approach maintains character recognition and interaction quality while significantly reducing the computational burden of rendering large numbers of characters.
2Manufacturing precision
If high-end computing devices are used to render all virtual characters in good quality, then rendering quality is maintained, but rendering resources are excessively consumed
Solution Approach 1:
The patent implements dynamic level of detail adjustment where the rendering quality of each virtual character is automatically adjusted in real-time based on multiple factors including distance from the camera, character importance, interaction state, and current device performance. This dynamic adaptation allows the system to maintain high rendering quality for critical characters while automatically reducing detail for less important ones, optimizing the balance between visual quality and computational resource consumption without requiring manual intervention.
Solution Approach 2:
The patent changes geometric parameters of character models by switching between pre-defined level of detail versions with varying mesh complexity. Each LOD version has different triangle counts and geometric details, allowing the renderer to select appropriate parameter sets based on current rendering needs. This parameter-based approach enables smooth quality adjustments and efficient resource allocation across multiple characters simultaneously.
3Ease of manufacture
If low-end computing devices are used, then device accessibility is improved, but real-time rendering capability is insufficient for multiple virtual characters
Solution Approach 1:
The patent applies partial action by rendering only the necessary portions of virtual characters at high quality while using simplified representations for the rest. Instead of rendering complete high-fidelity models for all characters, the system selectively applies detailed rendering only to visible and interaction-critical regions, allowing low-end devices to handle more characters in real-time by reducing the total geometric complexity that must be processed each frame.
Data Source
AI summary
A real-time rendering generating apparatus, method, and non-transitory computer readable storage medium thereof are provided. The apparatus receives a plurality of character motion data of a plurality of virtual characters. The apparatus determines a rendering level corresponding to each of the virtual characters based on a classification rule related to a first virtual character and the character motion data, each of the rendering levels corresponds to one of a plurality of character level of detail, and each of the plurality of character level of detail corresponds to a range of a customized body part and a skeletal model. The apparatus generates a real-time rendering of each of the virtual characters based on the rendering level corresponding to each of the virtual characters.