Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
GNeRF: GAN-based Neural Radiance Field without ...
Search
peisuke
December 12, 2021
Technology
1
790
GNeRF: GAN-based Neural Radiance Field without Posed Camera
第9回全日本コンピュータビジョン勉強会
ICCV2021論文読み回発表資料
GNeRF: GAN-based Neural Radiance Field without Posed Camera
peisuke
December 12, 2021
Tweet
Share
More Decks by peisuke
See All by peisuke
VGGT: Visual Geometry Grounded Transformer
peisuke
0
9
AI for Kids:小学生に画像認識を教えてみた話
peisuke
1
34
LangGraphで始めるマルチエージェントシステム
peisuke
13
4.5k
Self-RAG: Learning to Retrieve, Generate and Critique through Self-Reflections
peisuke
9
1.5k
Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields
peisuke
0
12k
LangChain Toolsの運用と改善
peisuke
5
2.8k
TTS Skins: Speaker Conversion via ASR
peisuke
0
410
A Quantum Computational Approach to Correspondence Problems on Point Sets
peisuke
0
730
F0-Consistent Many-to-many Non-parallel Voice Conversion via Conditional Autoencoder
peisuke
0
220
Other Decks in Technology
See All in Technology
Getting to Know Your Legacy (System) with AI-Driven Software Archeology (WeAreDevelopers World Congress 2025)
feststelltaste
1
160
20250705 Headlamp: 專注可擴展性的 Kubernetes 用戶界面
pichuang
0
280
Contributing to Rails? Start with the Gems You Already Use
yahonda
2
100
NewSQLや分散データベースを支えるRaftの仕組み - 仕組みを理解して知る得意不得意
hacomono
PRO
3
180
american airlines®️ USA Contact Numbers: Complete 2025 Support Guide
supportflight
1
110
OpenTelemetryセマンティック規約の恩恵とMackerel APMにおける活用例 / SRE NEXT 2025
mackerelio
2
360
Lufthansa ®️ USA Contact Numbers: Complete 2025 Support Guide
lufthanahelpsupport
0
210
AWS認定を取る中で感じたこと
siromi
1
190
[ JAWS-UG千葉支部 x 彩の国埼玉支部 ]ムダ遣い卒業!FinOpsで始めるAWSコスト最適化の第一歩
sh_fk2
2
110
DBのスキルで生き残る技術 - AI時代におけるテーブル設計の勘所
soudai
PRO
54
20k
What’s new in Android development tools
yanzm
0
340
United Airlines Customer Service– Call 1-833-341-3142 Now!
airhelp
0
170
Featured
See All Featured
Principles of Awesome APIs and How to Build Them.
keavy
126
17k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
29
1.8k
Build your cross-platform service in a week with App Engine
jlugia
231
18k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
45
7.5k
GitHub's CSS Performance
jonrohan
1031
460k
Fashionably flexible responsive web design (full day workshop)
malarkey
407
66k
Java REST API Framework Comparison - PWX 2021
mraible
31
8.7k
Building Applications with DynamoDB
mza
95
6.5k
Unsuck your backbone
ammeep
671
58k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.7k
Improving Core Web Vitals using Speculation Rules API
sergeychernyshev
18
970
Statistics for Hackers
jakevdp
799
220k
Transcript
GNeRF: GAN-based Neural Radiance Field without Posed Camera @peisuke 9
ABEJA Twitter @peisuke Github https://github.com/peisuke Qiita https://qiita.com/peisuke SlideShare https://www.slideshare.net/FujimotoKeisuke
• • GNeRF: GAN-based Neural Radiance Field without Posed Camera
• • NeRF • • • GAN Radiance Field • RF
NeRF • / 5 •
• • (x,y,z)T (φ1,θ1)T (φ2,θ2)T (x,y,z) φ2,θ2 (x,y,z) φ1,θ1
NeRF • • •
NeRF • NeRF in the Wild • NeRF • iNeRF
• iNeRF NeRF in the Wild BP
条件 iNeRF NeRF-- IDR GNeRF(提案⼿法) 事前学習 必要 不要 不要 不要
粗いカメラ座標 必要 必要 必要 不要 前⽅のみの制約 不要 必要 不要 不要 マスク 不要 不要 必要 不要 • •
• • • • • GAN Radiance Field • RF
• Phase A Phase B
RF • D E 3
GAN Real • • Discriminator Real • Dynamic Patch Sampling
/ 16x16 • Dynamic Patch Sampling
GAN Fake • • NeRF 16x16 • G
• 6 𝑹 = 𝑟!! 𝑟!" 𝑟!# 𝑟"! 𝑟"" 𝑟"#
𝑟#! 𝑟#" 𝑟## 𝒓 = 𝑟!! 𝑟!" 𝑟!# 𝑟"! 𝑟"" 𝑟"# $ • SO(3) NN • On the Continuity of Rotation Representations in Neural Networks
• Inversion network • Static Patch Sampling • E •
Static Patch Sampling
G G D G
• • Inversion network • refinement
• Refinement
• • • • • Photometric loss
• • Generator • NeRF • Discriminator • Instance Normalization
Spectral Normalization • Inversion network • Vision Transformer • • A • A B 1 iteration • B
• • COLMAP NeRF C+n • Refine C+r • Ground
Truth NeRF G+n • • Synthetic-NeRF image size: 400x400 DTU 500x400 • 100 • • Peak Signal-to-Noise Ratio SSIM LPIPS • / • Single TAITAN GPU 30
COLMAP •
• COLMAP Drums
難しいシーン 特徴の少ない シーン
IDR • 3D IDR • IDR • PNSR •
• COLMAP • Drums Lego COLMAP
Ablation Study • • GAN InvNet. PhaseB • • A
B A ABABABAB B
• • • COLMAP • •
• NeRF • • GAN RF • RF • •
• COLMAP