Our CGF paper "Time-Warped Foveated Rendering for VR Headsets" is now published with open access: doi.org/10.1111/cgf.14… We present foveated spatio-temporal rendering with a novel perceptual evaluation metric for reprojection quality resulting in up to 60% speedup

We're excited to present our paper "ADOP: Approximate Differentiable One-Pixel Point Rendering" at ACM SIGGRAPH 2022! Come by Tuesday at 4pm for the presentation and Thursday at 12:20 pm for the lab demo session. Project Page and Code: github.com/darglein/adop (including new scenes)

Introducing our paper "VET: Visual Error Tomography for Point Cloud Completion and High-Quality Neural Rendering" (SIGGRAPH Asia ➡️ Hong Kong 2023). Completing point clouds by lifting 2D visual error maps for photo realistic rendering. 📖: lfranke.github.io/vet 📜: arxiv.org/abs/2311.04634

If you are at SIGGRAPH Asia ➡️ Hong Kong, come see the presentation in room C4.11 at 4:40pm today!

FAU Erlangen-Nürnberg presents TRIPS Trilinear Point Splatting for Real-Time Radiance Field Rendering paper page: huggingface.co/papers/2401.06… Point-based radiance field rendering has demonstrated impressive results for novel view synthesis, offering a compelling blend of rendering

Excited to share our new paper! TRIPS: Trilinear Point Splatting for Real-Time Radiance Field Rendering tl;dr: Smooth point rendering via trilinear writes in an image pyramid for highly detailed radiance field renderings. 📖: lfranke.github.io/trips/ 📜: arxiv.org/abs/2401.06003

We have released the source code for TRIPS at github.com/lfranke/TRIPS Try it out! :)

Check out our latest work: Implicit Neural Point Clouds Volume-based scene reconstruction combined with point rendering and trained end-to-end for interactive radiance field rendering with super fine details. 📖: fhahlbohm.github.io/inpc/ Led by Florian Hahlbohm!

Thank you VCAI - MPI for Informatics for having me over at MPI in Saarbrücken! Great days presenting my research and talking to so many smart and enthusiastic people about neural point rendering!

Our latest work to increase multi-view coherence in 3D Gaussian Splatting using Hybrid Transparency: - Perspective-correct (no affine approximation) and without matrix inversions - Per-pixel sorted (no popping) for the K-nearest fragments ~2x faster than 3DGS Led by Florian Hahlbohm

The source code for our Hybrid Transparency Gaussian Splatting has been released! github.com/nerficg-projec… (Including a colab for the perspective-correct math part) Try it out! :) tl;dr: Perspective correct & no popping & ~2x faster than 3DGS

Anyone trying out our hybrid transparency Gaussian splatting ( fhahlbohm.github.io/htgs ) code: github.com/nerficg-projec… Take a look at this thread from Florian Hahlbohm with a lot of interesting details!