Gaze-Based Polygon Reduction for Low-Latency Shared XR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The significant processing load in displaying polygon data for virtual objects in XR systems can lead to delays, especially when multiple users are participating in a shared virtual space.
Innovation Solution
An information processing apparatus that specifies user attention positions, extracts non-attention objects, and reduces data for these objects based on attention criteria, using a data reduction unit to optimize data delivery to user terminals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution polygon data is displayed for all virtual objects, then display quality is improved, but processing load increases causing display delays
Solution Approach 1:
The patent applies local quality by differentiating display resolution based on user attention. Virtual objects in the attention region (where users are looking) are rendered at high resolution, while virtual objects in non-attention regions are rendered at low resolution. This selective quality adjustment reduces overall processing load while maintaining display quality where users actually perceive it, thereby resolving the contradiction between display quality and display delay.
2Productivity
If data reduction is applied to all virtual objects, then processing load is reduced, but display quality deteriorates
Solution Approach 1:
The patent implements local quality by applying data reduction selectively only to virtual objects in non-attention regions, while maintaining full data quality for virtual objects in attention regions. The attention-position specifying unit identifies where users are looking, and the data reduction unit applies compression only to objects outside these regions. This ensures processing efficiency is improved without sacrificing display quality in areas users actually observe.
3Adaptability or versatility
If multiple users participate in shared virtual space, then system versatility is improved, but processing load increases causing display delays
Solution Approach 1:
The patent resolves the multi-user processing load issue by applying local quality to each user's perspective individually. The attention-position specifying unit determines attention regions for each user based on their viewing direction and position. The data reduction unit then processes virtual objects differently for each user based on their specific attention regions. This allows the system to support multiple users with full-quality rendering only where each user is actually looking, thereby maintaining versatility while reducing overall processing requirements and display delays.
Data Source
AI summary
Attention position specifying unit specifies attention positions to which users pay attention in a virtual space, based on detection results of users' lines of sight acquired by acquisition unit, and three-dimensional coordinates in the virtual space displayed on head-mounted displays of the users at the time of detection of the lines of sight. Non-attention object extraction unit extracts as a non-attention object, from virtual objects included in the virtual space, a virtual object for which a number of times the virtual object is specified as an attention position does not meet a criterion. Data reduction unit reduces an amount of polygon data for displaying the non-attention object extracted by non-attention object extraction unit.