update abstract

fraunhoferhhi · May 14, 2024 · 27e2671 · 27e2671
1 parent 71fefbe
commit 27e2671
Showing 1 changed file with 24 additions and 19 deletions.
diff --git a/project-page/index.html b/project-page/index.html
@@ -3,19 +3,23 @@
 
 <head>
     <meta charset="utf-8"/>
-    <meta name="description" content="Self-Organizing Gaussian Splats: A new approach to 3D scene representation and rendering, reducing data size significantly while preserving quality, developed by researchers at Fraunhofer Heinrich Hertz Institute HHI and Humboldt University of Berlin."/>
+    <meta name="description"
+          content="Self-Organizing Gaussian Splats: A new approach to 3D scene representation and rendering, reducing data size significantly while preserving quality, developed by researchers at Fraunhofer Heinrich Hertz Institute HHI and Humboldt University of Berlin."/>
     <meta property="og:title" content="Self-Organizing Gaussian Splats: Efficient 3D Scene Representation"/>
-    <meta property="og:description" content="Introducing Self-Organizing Gaussian Splats for compact 3D scene representation. This method significantly reduces the data size needed for high-quality 3D rendering, ideal for applications on resource-constrained devices and fast web applications."/>
+    <meta property="og:description"
+          content="Introducing Self-Organizing Gaussian Splats for compact 3D scene representation. This method significantly reduces the data size needed for high-quality 3D rendering, ideal for applications on resource-constrained devices and fast web applications."/>
     <meta property="og:url" content="https://fraunhoferhhi.github.io/Self-Organizing-Gaussian-Splats/"/>
     <meta property="og:image" content="static/images/teaser_self_organizing_gaussian.jpeg"/>
     <meta property="og:image:width" content="1200"/>
     <meta property="og:image:height" content="675"/>
 
     <meta name="twitter:title" content="Discover Self-Organizing Gaussian Splats for 3D Scene Representation"/>
-    <meta name="twitter:description" content="Learn about the novel Self-Organizing Gaussian Splats method, a groundbreaking approach in 3D scene rendering, offering high-quality visuals with reduced data size."/>
+    <meta name="twitter:description"
+          content="Learn about the novel Self-Organizing Gaussian Splats method, a groundbreaking approach in 3D scene rendering, offering high-quality visuals with reduced data size."/>
     <meta name="twitter:image" content="static/images/your_twitter_banner_image.png"/>
     <meta name="twitter:card" content="summary_large_image"/>
-    <meta name="keywords" content="Self-Organizing Gaussian Splats, 3D scene representation, compact data size, high-quality rendering, Fraunhofer Heinrich Hertz Institute, Humboldt University of Berlin, resource-constrained devices, fast web applications"/>
+    <meta name="keywords"
+          content="Self-Organizing Gaussian Splats, 3D scene representation, compact data size, high-quality rendering, Fraunhofer Heinrich Hertz Institute, Humboldt University of Berlin, resource-constrained devices, fast web applications"/>
     <meta name="viewport" content="width=device-width, initial-scale=1"/>
 
     <title>Self-Organizing Gaussian Splats Project Page</title>
@@ -134,21 +138,22 @@ <h2 class="subtitle has-text-centered">
                 <h2 class="title is-3">Abstract</h2>
                 <div class="content has-text-justified">
                     <p>
-                        3D Gaussian Splatting has recently emerged as a highly promising technique for modeling of static 3D
-                        scenes. In contrast to Neural Radiance Fields, it utilizes efficient rasterization allowing for very fast
-                        rendering at high-quality. However, the storage size is significantly higher, which hinders practical
-                        deployment, e.g.on resource constrained devices. In this paper, we introduce a compact scene
-                        representation organizing the parameters of 3D Gaussian Splatting (3DGS) into a 2D grid with local
-                        homogeneity, ensuring a drastic reduction in storage requirements without compromising visual quality
-                        during rendering. Central to our idea is the explicit exploitation of perceptual redundancies present in
-                        natural scenes. In essence, the inherent nature of a scene allows for numerous permutations of Gaussian
-                        parameters to equivalently represent it. To this end, we propose a novel highly parallel algorithm that
-                        regularly arranges the high-dimensional Gaussian parameters into a 2D grid while preserving their
-                        neighborhood structure. During training, we further enforce local smoothness between the sorted parameters
-                        in the grid. The uncompressed Gaussians use the same structure as 3DGS, ensuring a seamless integration
-                        with established renderers. Our method achieves a reduction factor of 8x to 26x in size for complex scenes
-                        with no increase in training time, marking a substantial leap forward in the domain of 3D scene
-                        distribution and consumption.
+                        3D Gaussian Splatting has recently emerged as a highly promising technique for modeling of
+                        static 3D scenes. In contrast to Neural Radiance Fields, it utilizes efficient rasterization
+                        allowing for very fast rendering at high-quality. However, the storage size is significantly
+                        higher, which hinders practical deployment, e.g.~on resource constrained devices. In this paper,
+                        we introduce a compact scene representation organizing the parameters of 3D Gaussian Splatting
+                        (3DGS) into a 2D grid with local homogeneity, ensuring a drastic reduction in storage
+                        requirements without compromising visual quality during rendering.
+                        Central to our idea is the explicit exploitation of perceptual redundancies present in natural
+                        scenes. In essence, the inherent nature of a scene allows for numerous permutations of Gaussian
+                        parameters to equivalently represent it. To this end, we propose a novel highly parallel
+                        algorithm that regularly arranges the high-dimensional Gaussian parameters into a 2D grid while
+                        preserving their neighborhood structure. During training, we further enforce local smoothness
+                        between the sorted parameters in the grid. The uncompressed Gaussians use the same structure as
+                        3DGS, ensuring a seamless integration with established renderers. Our method achieves a
+                        reduction factor of 17x to 42x in size for complex scenes with no increase in training time,
+                        marking a substantial leap forward in the domain of 3D scene distribution and consumption.
                     </p>
                 </div>
             </div>