PowerGS: Display-Rendering Power Co-Optimization for Neural Rendering in Power-Constrained XR Systems

2025-09-25

AI Summary

To address the excessive rendering and display power consumption of 3D Gaussian Splatting (3DGS) on extended reality (XR) devices under watt-level power constraints, this paper proposes the first framework jointly optimizing neural rendering and display power. Our method introduces a differentiable rendering-display co-power model, incorporates nonlinear power modeling under image quality constraints, and achieves closed-form global power minimization via iso-quality curves. We further integrate foveated rendering to enhance energy efficiency. Experimental results demonstrate that, with negligible subjective and objective quality degradation, the total system power consumption is reduced by up to 86% compared to state-of-the-art methods. This significant reduction markedly improves the energy efficiency ratio and battery longevity of XR devices, enabling practical deployment of high-fidelity 3DGS rendering under stringent power budgets.

Technology Category

• Computer Vision: 3D Computer Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Non-convex Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

3D Gaussian Splatting (3DGS) combines classic image-based rendering, pointbased graphics, and modern differentiable techniques, and offers an interesting alternative to traditional physically-based rendering. 3DGS-family models are far from efficient for power-constrained Extended Reality (XR) devices, which need to operate at a Watt-level. This paper introduces PowerGS, the first framework to jointly minimize the rendering and display power in 3DGS under a quality constraint. We present a general problem formulation and show that solving the problem amounts to 1) identifying the iso-quality curve(s) in the landscape subtended by the display and rendering power and 2) identifying the power-minimal point on a given curve, which has a closed-form solution given a proper parameterization of the curves. PowerGS also readily supports foveated rendering for further power savings. Extensive experiments and user studies show that PowerGS achieves up to 86% total power reduction compared to state-of-the-art 3DGS models, with minimal loss in both subjective and objective quality. Code is available at https://github.com/horizon-research/PowerGS.

Problem

• Optimizing rendering and display power for 3D Gaussian Splatting in XR devices
• Minimizing total power consumption under quality constraints for neural rendering
• Enabling foveated rendering to achieve significant power savings in XR systems

Innovation

• Jointly minimizes rendering and display power consumption
• Identifies iso-quality curves for power optimization
• Supports foveated rendering for additional power savings