NEAT: A Rendering-Distilling Approach for Efficient 3D Wireframe Reconstruction from Multi-View Images
This paper presents NEAT, a novel rendering-distilling formulation using neural fields to represent 3D line segments and junctions, enabling matching-free 3D wireframe reconstruction from multi-view images.
GD2-NeRF: Generative Detail Compensation via GAN and Diffusion for One-shot Generalizable Neural Radiance Fields
GD2-NeRF is a coarse-to-fine generative detail compensation framework that hierarchically includes GAN and pre-trained diffusion models into One-shot Generalizable Neural Radiance Fields (OG-NeRF) to synthesize novel views with vivid plausible details in an inference-time finetuning-free manner.
NeRF-based 3D Representation Disentanglement via Latent Semantic Navigation
NaviNeRF, a NeRF-based 3D reconstruction model, achieves fine-grained disentanglement while preserving 3D accuracy and consistency without any priors or supervision.
Coherent 3D Gaussian Splatting for Sparse Novel View Synthesis
We propose a regularized optimization approach to enable 3D Gaussian Splatting (3DGS) for sparse input views. Our key idea is to introduce coherency to the 3D Gaussians during optimization by constraining their movement in 2D image space using an implicit decoder and total variation loss. We further leverage monocular depth and flow correspondences to initialize and regularize the 3D Gaussian representation, enabling high-quality texture and geometry reconstruction from extremely sparse inputs.
Decomposed 3D Scene Reconstruction with Minimal Human Interaction
A novel method for decomposing 3D scenes into individual objects and backgrounds with minimal human interaction, by integrating the Segment Anything Model (SAM) with hybrid implicit-explicit neural surface representations and a mesh-based region-growing technique.
© 2024 by Linnk AI