Neural Surface Reconstruction (NeuS) vs. Photogrammetry

preview_player
Показать описание


Happy Halloween!
  1. Matthew Brennan
Рекомендации по теме
Neural Surface Reconstruction 0:12:38

Neural Surface Reconstruction (NeuS) vs. Photogrammetry

NeuS2: Fast Learning 0:05:38

NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction. In ICCV, 2023.

Recovering Fine Details 0:03:45

Recovering Fine Details for Neural Implicit Surface Reconstruction

Ref-NeuS: Ambiguity-Reduced Neural 0:04:59

Ref-NeuS: Ambiguity-Reduced Neural Implicit Surface Learning for Multi-View Reconstruction with Ref

PET-NeuS: Positional Encoding 0:06:01

PET-NeuS: Positional Encoding Tri-Planes for Neural Surfaces

LightNeuS: Neural Surface 0:05:01

LightNeuS: Neural Surface Reconstruction in Endoscopy using Illumination Decline

NeuS2: Fast Learning 0:05:01

NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction

Neural Surface Reconstruction 0:07:01

Neural Surface Reconstruction Against High Specular Reflections (CVPR'2023)

NLOS-NeuS: Non-line-of-sight Neural 0:04:43

NLOS-NeuS: Non-line-of-sight Neural Implicit Surface

High-Fidelity Mask-free Neural 0:05:05

High-Fidelity Mask-free Neural Surface Reconstruction for Virtual Reality

NeUDF: Leaning Neural 0:07:59

NeUDF: Leaning Neural Unsigned Distance Fields with Volume Rendering

[CVPR 2023] Towards 0:07:18

[CVPR 2023] Towards Unbiased Volume Rendering of Neural Implicit Surfaces With Geometry Priors

[ECCV 2024] PS-NEUS: 0:04:50

[ECCV 2024] PS-NEUS: A Probability-guided Sampler for Neural Implicit Surface Rendering

[CVPR 2023] Neuralangelo: 0:06:26

[CVPR 2023] Neuralangelo: High-Fidelity Neural Surface Reconstruction

[CVPR 2023] NeAT: 0:08:01

[CVPR 2023] NeAT: Learning Neural Implicit Surfaces with Arbitrary Topologies from Multi-view Images

CSC2547 DeepSDF 0:10:07

CSC2547 DeepSDF Learning Continuous Signed Distance Functions for Shape Representation

NVIDIA’s Neuralangelo AI: 0:05:08

NVIDIA’s Neuralangelo AI: Gaming Anywhere on Earth!

Why is My 0:00:59

Why is My Featureless Object Scan Result Bad in KIRI Engine #faq #tutorial #neus #3dscan #kiriengine

Introduction to Featureless 0:00:59

Introduction to Featureless Object Scan in KIRI Engine #computergraphics #ai #neus #nsr #3dscan

S-VolSDF: Sparse Multi-View 0:04:50

S-VolSDF: Sparse Multi-View Stereo Regularization of Neural Implicit Surfaces

Neural 3D Scene 0:05:15

Neural 3D Scene Reconstruction With the Manhattan World Assumption | CVPR 2022

Video to 3D 0:01:01

Video to 3D by Nvidia AI - Nuralangelo #archivlogs #architecture #ai #nvidia

Nvidia's Neuralangelo: Transforming 0:00:46

Nvidia's Neuralangelo: Transforming 2D Videos into Stunning 3D

3DGV Talk: Wenping 1:26:40

3DGV Talk: Wenping Wang --- Studies on 3D Reconstruction.

Комментарии
Автор

Another great video on this topic, thanks! I capture a lot of 3d scans for museums, and I have had to capture several super reflective or shiny featureless objects. With priceless museum objects you can’t use the usual talc powder or sprays to coat the objects or use marker stickers for better photogrammetry results, so something like this would be a huge help for those types of objects if the quality can improve as they develop this!

edwardrobison
Автор

Create video,
I also tested KIRI Engine two months ago, and I found the featureless mode really useful. For example, with the plastic toy figures which do have a lot of reflective mono-colored surfaces. The only issue I had then was that the scans sometimes got cropped on the bottom or the top I guess that has to do with the aspect ratio that KIRI Engine is expected
Anyway, really enjoying your videos. 🎉

buroachenbach
Автор

Neus is ok if your object can't be processed using photogrammetry like transparent or translucent objects. I had spent 30 hours trying to construct a figure. It works ok but the level of detail is terrible.

michaelwhite
Автор

Someone should turn Google Street View into 3d. It would be so nice to drive around and explore the world in a driving game.

jimj