SWAG: Splatting in the Wild images with Appearance-conditioned Gaussians
Keywords: 3D Gaussian Splatting · Unconstrained Photo Collection, Novel View Synthesis, Appearance Modeling, Real-time Rendering, Transient Object Removal
Abstract: Implicit neural representation methods have shown impres-
sive advancements in learning 3D scenes from unstructured in-the-wild
photo collections but are still limited by the large computational cost
of volumetric rendering. Recently, 3D Gaussian Splatting emerged as
a much faster alternative with superior rendering quality and training
efficiency, especially for small-scale and object-centric scenarios. Never-
theless, this technique suffers from poor performance on unstructured
in-the-wild data. To tackle this, we extend over 3D Gaussian Splatting
to handle unstructured image collections. We achieve this by modeling
appearance to seize photometric variations in the rendered images. Ad-
ditionally, we introduce a new mechanism to train transient Gaussians to
handle the presence of scene occluders in an unsupervised manner. Ex-
periments on diverse photo collection scenes and multi-pass acquisition
of outdoor landmarks show the effectiveness of our method over prior
works achieving state-of-the-art results with improved efficiency.
Submission Number: 3
Loading