From 0e5eeb366fa42e443aa1089518297ecba17df460 Mon Sep 17 00:00:00 2001 From: cruc <93235593+tongji-rkr@users.noreply.github.com> Date: Tue, 26 Mar 2024 22:27:32 +0800 Subject: [PATCH] ver3.1 --- octree-gs/index.html | 22 ++++++---------------- 1 file changed, 6 insertions(+), 16 deletions(-) diff --git a/octree-gs/index.html b/octree-gs/index.html index 5a475f5..b110b57 100644 --- a/octree-gs/index.html +++ b/octree-gs/index.html @@ -174,7 +174,7 @@

Octree-GS: Towards Consistent Real-time --> - @@ -182,16 +182,6 @@

Octree-GS: Towards Consistent Real-time Code - - - - - - - Viewer - - @@ -206,7 +196,7 @@

Octree-GS: Towards Consistent Real-time

- TL;DR: We introduce Octree-GS, featuring an LOD-structured 3D Gaussian approach supporting level-of-detail decomposition for both scene geometry and rendering results. + TL;DR: We introduce Octree-GS, featuring an LOD-structured 3D Gaussian approach supporting level-of-detail decomposition for scene representation that contributes to the final rendering results.

@@ -232,11 +222,11 @@

Octree-GS: Towards Consistent Real-time

Abstract

- The recent 3D Gaussian splatting (3D-GS) has shown remarkable rendering fidelity and excels in speed compared to NeRF-basedneural scene representations. - While demonstrating the potential for realtime rendering, 3D-GS encounters rendering bottlenecks in large scenes with complex details due to an excessive number of Gaussian primitives located within the viewing frustum. + The recent 3D Gaussian splatting (3D-GS) has shown remarkable rendering fidelity and efficiency compared to NeRF-based neural scene representations. + While demonstrating the potential for real-time rendering, 3D-GS encounters rendering bottlenecks in large scenes with complex details due to an excessive number of Gaussian primitives located within the viewing frustum. This limitation is particularly noticeable in zoom-out views and can lead to inconsistent rendering speeds in scenes with varying details. - Moreover, it often struggles to capture the corresponding level of details to the different scales with its heuristic Gaussian growing operation. - Inspired by the Level-of-details (LOD) techniques, we introduce Octree-GS, featuring an LOD-structured 3D Gaussian approach supporting level-of-detail decomposition for both scene geometry and rendering results. + Moreover, it often struggles to capture the corresponding level of details at different scales with its heuristic density control operation. + Inspired by the Level-of-Detail (LOD) techniques, we introduce \modelname, featuring an LOD-structured 3D Gaussian approach supporting level-of-detail decomposition for scene representation that contributes to the final rendering results. Our model dynamically selects the appropriate level from the set of multi-resolution anchor points, ensuring consistent rendering performance with adaptive LOD adjustments while maintaining high-fidelity rendering results.