Feature Splatting for Better Novel View Synthesis with Low Overlap

TL;DR

This paper introduces Feature Splatting, a new method that encodes scene colors into feature vectors for 3D Gaussian Splatting, significantly improving the quality of novel view synthesis especially with low overlap views.

Contribution

The paper proposes a novel feature vector encoding for 3D Gaussians, enhancing generalization and rendering quality in novel view synthesis beyond spherical harmonics limitations.

Findings

Improved view synthesis quality for distant, low-overlap views.

Capable of generating per-pixel semantic labels.

Enhanced expressivity over traditional spherical harmonics representation.

Abstract

3D Gaussian Splatting has emerged as a very promising scene representation, achieving state-of-the-art quality in novel view synthesis significantly faster than competing alternatives. However, its use of spherical harmonics to represent scene colors limits the expressivity of 3D Gaussians and, as a consequence, the capability of the representation to generalize as we move away from the training views. In this paper, we propose to encode the color information of 3D Gaussians into per-Gaussian feature vectors, which we denote as Feature Splatting (FeatSplat). To synthesize a novel view, Gaussians are first "splatted" into the image plane, then the corresponding feature vectors are alpha-blended, and finally the blended vector is decoded by a small MLP to render the RGB pixel values. To further inform the model, we concatenate a camera embedding to the blended feature vector, to condition the decoding also on the viewpoint information. Our experiments show that these novel model for encoding the radiance considerably improves novel view synthesis for low overlap views that are distant from the training views. Finally, we also show the capacity and convenience of our feature vector representation, demonstrating its capability not only to generate RGB values for novel views, but also their per-pixel semantic labels. Code available at https://github.com/tberriel/FeatSplat . Keywords: Gaussian Splatting, Novel View Synthesis, Feature Splatting