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
三次元形状とディープラーニング
Search
Tatsuya Yatagawa
December 21, 2020
Technology
310
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
三次元形状とディープラーニング
三次元形状処理とディープラーニングの初歩についてまとめたスライドです。
Tatsuya Yatagawa
December 21, 2020
More Decks by Tatsuya Yatagawa
See All by Tatsuya Yatagawa
Event-based Camera Simulation using Monte Carlo Path Tracing with Adaptive Denoising (ICIP 2023)
tatsy
0
590
VC'23 チュートリアル講演 『三次元メッシュと深層学習』
tatsy
0
240
SIGGRAPH 2022: Neural Geometry Processing
tatsy
0
260
Real-time Shading with Free-form Planar Area Lights using Linearly Transformed Cosines (JCGT & I3D 2022)
tatsy
0
1.1k
SIGGRAPH Asia 2021: Real-Time Rendering
tatsy
0
160
SIGGRAPH 2021: Denoising and Guiding
tatsy
0
120
SIGGRAPH 2021: Shape Generation
tatsy
0
200
SIGGRAPH 2021: Meshing
tatsy
0
130
12. Keyboard & Mouse
tatsy
0
240
Other Decks in Technology
See All in Technology
システム思考で問題に対処する
yussak
0
240
【CEDEC2026】専門性の高いデフォルメチームが挑んだ人材育成戦略 〜Cygames Academiaの企画から実施まで〜
cygames
PRO
0
570
JavaScript 研修 (2026)
recruitengineers
PRO
2
520
老害フォレンジッカーはAI羊の夢を見るか?
tadmaddad
0
320
Bill One 開発エンジニア 紹介資料
sansan33
PRO
7
19k
サイバー捜査員研修(前半)
nomizone
1
2k
モバイルアプリ開発概論2026
recruitengineers
PRO
5
560
ブラウザ研修 2026
recruitengineers
PRO
6
930
Eight Engineering Unit 紹介資料
sansan33
PRO
3
8.2k
名刺メーカーDevグループ 紹介資料
sansan33
PRO
0
1.2k
FPGAが実現する遠方宇宙の高空間分解能天体撮影 -大型地上望遠鏡の視力を補正する「補償光学」とは?-
komei_mt
0
220
【CEDEC2026】コードレビュー支援ツール開発から学ぶ:LLMを用いた業務システムの実践的な運用設計と誤出力対策
cygames
PRO
0
660
Featured
See All Featured
Deep Space Network (abreviated)
tonyrice
0
250
We Are The Robots
honzajavorek
0
300
The Hidden Cost of Media on the Web [PixelPalooza 2025]
tammyeverts
2
460
We Analyzed 250 Million AI Search Results: Here's What I Found
joshbly
1
1.7k
Data-driven link building: lessons from a $708K investment (BrightonSEO talk)
szymonslowik
1
1.2k
Tips & Tricks on How to Get Your First Job In Tech
honzajavorek
1
690
Docker and Python
trallard
47
4.1k
Lightning talk: Run Django tests with GitHub Actions
sabderemane
0
230
How People are Using Generative and Agentic AI to Supercharge Their Products, Projects, Services and Value Streams Today
helenjbeal
1
260
Designing Experiences People Love
moore
143
24k
Leveraging Curiosity to Care for An Aging Population
cassininazir
1
460
Reality Check: Gamification 10 Years Later
codingconduct
0
2.2k
Transcript
None
• 機械学習 ニューラルネット 深層学習
• • • θ • = ; = 2 2
+ 1 + 0 ෩ = arg min =0 −1 ( , ; )
• • • 0 = 0 + 0 1 =
1 + 1 2 = 2 + 2 = 0 0 + 1 1 + 2 2 + = 0 0 + 1 1 + 2 2 +(0 0 + 1 1 + 2 2 + )
• • • 0 1 2 0 = ReLU(0 +
0 ) 1 = ReLU(1 + 1 ) 2 = ReLU(2 + 2 )
= 32 = 128 = 512 • •
• • = 512 = 1 = = 4
•
•
•
•
•
•
• • • • • • • • • •
•
(c) Velodyne (c) Andrew Tallon (c) Microsoft (c) Sony
• , , = 0
′ = − ′ = + ′
• ☺ ☺ ☺ ☺ ,
, = 0
☺ ☺ ☺ ☺ , ,
= 0
• • • • •
• • • •
•
•
• ⚫ ⚫ ⚫
• • •
• • •
• •
• • •
• • • • • • •
☺ ☺ ☺ ☺ , ,
= 0
• • • (c) Velodyne
• • • (0 , 0 , 0 ) (1
, 1 , 1 ) ( , , ) (−1 , −1 , −1 ) (+1 , +1 , +1 ) (0 , 0 , 0 ) (1 , 1 , 1 ) (−1 , −1 , −1 ) (+1 , +1 , +1 ) ( , , )
• • •
None
⚫
⚫
• •
• • •
•
•
•
• • • • • • •
☺ ☺ ☺ ☺ , ,
= 0
• • •
• • • •
• • ∗ = ℱ−1(ℱ ⋅ ℱ ) ′ =
ℱ−1(Θ ∘ ℱ )
• = 0 1 1 0 1 0 1 0
1 1 0 1 0 0 1 0 = − − 1 2− 1 2 = :
• • • = T
• () ′ = ℱ−1(Θ ∘ ℱ ) ′ =
ℱ = () ℱ−1 = ()
• • • +1 +2 +1 = ReLU ⋅
• • •
• • •
• •
• • •
• • •
• •
• • • • • •
☺ ☺ ☺ ☺ , ,
= 0
• , , =
• • •
• • • •
• •
• • •
• • •
• • • •
•
• • • •
• • •
• • • •
• • •
• • • •
• • • • • • • •
• • • • • • • • •