VGGT (Visual Geometry Grounded Transformer) leverages deep learning based representations to infer 3D structures from an image rather than traditional 2D based SfM pipelines. It provides a simplified, ...
MASt3R and MASt3R-SfM Explanation: Image Matching and 3D Reconstruction Results
MASt3R (Matching and Stereo 3D Reconstruction) aims to treat image matching as a 3D problem leveraging dense correspondences and understanding the 3D scene rather than a traditional 2D approach. This ...
DUSt3R: Geometric 3D Vision Made Easy : Explanation and Results
DUSt3R (Dense and Unconstrained Stereo 3D Reconstruction) introduces a novel paradigm in multi-view 3D reconstruction, eliminating the need for predefined camera poses and intrinsics. 3D ...
Object Insertion in Gaussian Splatting: Paper Explanation and Training of MCMC in Gsplat
3D Gaussian splatting (3DGS) has recently gained recognition as a groundbreaking approach in radiance fields and computer graphics. It stands out as a jack of all trades, addressing challenges that ...
3D Gaussian Splatting Introduction – Paper Explanation & Training on Custom Datasets with NeRF Studio Gsplats
3D Gaussian Splatting (3DGS) is redefining the landscape of 3D computer graphics and vision — but here’s a twist: it achieves groundbreaking results without relying on any neural networks, not even a ...
The Annotated NeRF – Training on Custom Dataset from Scratch in Pytorch
In recent years, the field of 3D from multi-view has become one of the most popular topics in computer vision conferences, with a high number of submitted papers each year. A groundbreaking paper in ...
