Abstract Our goal is to reconstruct scenes with stochastic, incoherent motion such as leaves moving in the wind, that can be particularly challenging because of small objects with similar appearance that move independently. Previous dynamic 3D Gaussian Splatting solutions either represent motion implicitly with neural networks achieving good quality but lower framerate, or explicitly with a function, often with higher training times and lower quality. To overcome these limitations, we propose an explicit method that introduces adaptive space‐time densification and smoother optimization. We introduce a new densification approach based on error moments that are used to guide primitive splitting, and we adaptively refine the number of keyframes used based on the variance of error. We observe that dynamic reconstruction from monocular video is hard for standard optimization pipelines. To counter this, we introduce a weighted Adam approach that improves results based on primitive visibility. Finally, to handle the hard case of independent motion of similar‐looking objects, we introduce an image‐driven as‐rigid‐as‐possible regularization. Our method has higher quality than previous explicit solutions, and has significantly higher framerate for rendering.

Radiance fields such as 3D Gaussian Splatting allow real-time rendering of scenes captured from photos. They also reconstruct most specular reflections with high visual quality, but typically model them with “fake” reflected geometry, using primitives behind the reflector. Our goal is to correctly reconstruct the reflector and the reflected objects such as to make specular reflections editable; we present a proof of concept which exploits promising learning-based methods to extract diffuse and specular buffers from photos, as well as geometry and BRDF buffers. Our method builds on three key components. First, by using diffuse/specular buffers of input training views, we optimize a diffuse version of the scene and use path tracing to efficiently generate physically-based specular reflections. Second, we present a specialized training method that allows this process to converge. Finally, we present a fast ray tracing algorithm for 3D Gaussian primitives that enables efficient multi-bounce reflections. Our method reconstructs reflectors and reflected objects—including those not seen in the input images—in a unique scene representation. Our solution allows real-time, consistent editing of captured scenes with specular reflections, including multi-bounce effects, changing roughness etc. We mainly show results using ground truth buffers from synthetic scenes, and also preliminary results in real scenes with currently imperfect learning-based buffers. Code and data are available at: https://repo-sam.inria.fr/nerphys/editable-gaussian-reflections/.

gsplat is an open-source library designed for training and developing Gaussian Splatting methods. It features a front-end with Python bindings compatible with the PyTorch library and a back-end with highly optimized CUDA kernels. gsplat offers numerous features that enhance the optimization of Gaussian Splatting models, which include optimization improvements for speed, memory, and convergence times. Experimental results demonstrate that gsplat achieves up to 10% less training time and 4x less memory than the original implementation. Utilized in several research projects, gsplat is actively maintained on GitHub. Source code is available at https://github.com/nerfstudio-project/gsplat under Apache License 2.0. We welcome contributions from the open-source community.