InAcoustic ExplorationsbyJan Steckel·Oct 24How a bat could build a 3D model from its own echoesWhen you walk around a statue, you never see all of it at once, as certain parts are occluded by the object itself. Still, after a few…
Shanmuka Sadhu·Jun 19Paper Insights: PhysDreamer: Physics-Based Interaction with 3D Objects via Video GenerationThere has been recent research activity in the field of physical property estimation, and I wanted to read a few papers to become familiar…
Shanmuka Sadhu·Jun 18Paper Insights: Physical Property Understanding from Language-Embedded Feature FieldsThere has been recent research activity in the field of physical property estimation, and I wanted to read a few papers to become familiar…
Papers in 100 Lines of Code·Apr 13Fast 3D Reconstruction from a Single ImageUsing Meta-Learning to Initialize Neural Radiance Fields for Rapid, High-Quality Scene Generation in PyTorch
Papers in 100 Lines of Code·Mar 28NeRF vs 3D Gaussian Splatting: Why 3DGS Isn’t Better at EverythingA practical comparison of where 3DGS excels — and where NeRF still matters.
Shanmuka Sadhu·Mar 1Paper Insights: Neus: Learning Neural Implicit Surfaces by Volume Rendering for Multiview…Based on the many neural representation papers I have been reading, I believe the Neus paper was the next inevitable step. The idea of…
Shanmuka Sadhu·Mar 1Paper Insights: Block-NeRF: Scalable Large Scene Neural View SynthesisI recently watched a talk on why NeRFs are dead and why methods such as Gaussian Splatting are alternatives. However, in the talk, he gave…
Shanmuka Sadhu·Feb 22Triple Paper Insights: SIREN, WIRE, & FINERRecently, I started looking more deeply into neural fields and implicit neural representations (INR). Although I am a few years late to the…
Shanmuka Sadhu·Feb 17Paper Insights: NeRF: Representing Scenes as Neural Radiance Fields for View SynthesisCurrently, in the lab I am in, lab members are performing research on volumetric reconstruction and view synthesis. One of the most pivotal…
Papers in 100 Lines of Code·Feb 13Custom Gradients vs Autograd in 3D Gaussian SplattingIn a previous blog post, I showed how to write custom backward passes for 3D Gaussian Splatting (3DGS) -and, surprisingly, it was easy to…
