Shape, light, and material decomposition from images using monte carlo rendering and denoising

J Hasselgren, N Hofmann… - Advances in Neural …, 2022 - proceedings.neurips.cc
Recent advances in differentiable rendering have enabled high-quality reconstruction of 3D
scenes from multi-view images. Most methods rely on simple rendering algorithms: pre …

Towards better gradient consistency for neural signed distance functions via level set alignment

B Ma, J Zhou, YS Liu, Z Han - Proceedings of the IEEE/CVF …, 2023 - openaccess.thecvf.com
Neural signed distance functions (SDFs) have shown remarkable capability in representing
geometry with details. However, without signed distance supervision, it is still a challenge to …

Diff-lfd: Contact-aware model-based learning from visual demonstration for robotic manipulation via differentiable physics-based simulation and rendering

X Zhu, JH Ke, Z Xu, Z Sun, B Bai, J Lv… - … on Robot Learning, 2023 - proceedings.mlr.press
Abstract Learning from Demonstration (LfD) is an efficient technique for robots to acquire
new skills through expert observation, significantly mitigating the need for laborious manual …

Recursive control variates for inverse rendering

B Nicolet, F Rousselle, J Novak, A Keller… - ACM Transactions on …, 2023 - dl.acm.org
We present a method for reducing errors---variance and bias---in physically based
differentiable rendering (PBDR). Typical applications of PBDR repeatedly render a scene as …

Unsupervised inference of signed distance functions from single sparse point clouds without learning priors

C Chen, YS Liu, Z Han - … of the IEEE/CVF Conference on …, 2023 - openaccess.thecvf.com
It is vital to infer signed distance functions (SDFs) from 3D point clouds. The latest methods
rely on generalizing the priors learned from large scale supervision. However, the learned …

Gridpull: Towards scalability in learning implicit representations from 3d point clouds

C Chen, YS Liu, Z Han - Proceedings of the ieee/cvf …, 2023 - openaccess.thecvf.com
Learning implicit representations has been a widely used solution for surface reconstruction
from 3D point clouds. The latest methods infer a distance or occupancy field by overfitting a …

Learning neural implicit through volume rendering with attentive depth fusion priors

P Hu, Z Han - Advances in Neural Information Processing …, 2023 - proceedings.neurips.cc
Learning neural implicit representations has achieved remarkable performance in 3D
reconstruction from multi-view images. Current methods use volume rendering to render …

Coordinate quantized neural implicit representations for multi-view reconstruction

S Jiang, J Hua, Z Han - Proceedings of the IEEE/CVF …, 2023 - openaccess.thecvf.com
In recent years, huge progress has been made on learn-ing neural implicit representations
from multi-view images for 3D reconstruction. As an additional input complement-ing …

Differentiable rendering of neural sdfs through reparameterization

SP Bangaru, M Gharbi, F Luan, TM Li… - SIGGRAPH Asia 2022 …, 2022 - dl.acm.org
We present a method to automatically compute correct gradients with respect to geometric
scene parameters in neural SDF renderers. Recent physically-based differentiable …

Humans as light bulbs: 3d human reconstruction from thermal reflection

R Liu, C Vondrick - … of the IEEE/CVF Conference on …, 2023 - openaccess.thecvf.com
The relatively hot temperature of the human body causes people to turn into long-wave
infrared light sources. Since this emitted light has a larger wavelength than visible light …