$30 off During Our Annual Pro Sale. View Details »
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
0
280
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
Tweet
Share
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
510
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
200
SIGGRAPH 2022: Neural Geometry Processing
tatsy
0
220
Real-time Shading with Free-form Planar Area Lights using Linearly Transformed Cosines (JCGT & I3D 2022)
tatsy
0
950
SIGGRAPH Asia 2021: Real-Time Rendering
tatsy
0
100
SIGGRAPH 2021: Denoising and Guiding
tatsy
0
69
SIGGRAPH 2021: Shape Generation
tatsy
0
160
SIGGRAPH 2021: Meshing
tatsy
0
99
12. Keyboard & Mouse
tatsy
0
200
Other Decks in Technology
See All in Technology
AIにおける自由の追求
shujisado
3
470
一億総業務改善を支える社内AIエージェント基盤の要諦
yukukotani
8
2.8k
Contract One Engineering Unit 紹介資料
sansan33
PRO
0
9.9k
Oracle Database@AWS:サービス概要のご紹介
oracle4engineer
PRO
0
120
Introduction to Sansan, inc / Sansan Global Development Center, Inc.
sansan33
PRO
0
2.9k
All About Sansan – for New Global Engineers
sansan33
PRO
1
1.3k
名刺メーカーDevグループ 紹介資料
sansan33
PRO
0
980
AI/MLのマルチテナント基盤を支えるコンテナ技術
pfn
PRO
5
730
手動から自動へ、そしてその先へ
moritamasami
0
190
How native lazy objects will change Doctrine and Symfony forever
beberlei
1
380
会社紹介資料 / Sansan Company Profile
sansan33
PRO
11
390k
モバイルゲーム開発におけるエージェント技術活用への試行錯誤 ~開発効率化へのアプローチの紹介と未来に向けた展望~
qualiarts
0
300
Featured
See All Featured
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
46
7.8k
Being A Developer After 40
akosma
91
590k
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
36
6.2k
How Fast Is Fast Enough? [PerfNow 2025]
tammyeverts
3
380
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
12
1.3k
jQuery: Nuts, Bolts and Bling
dougneiner
65
8.1k
RailsConf 2023
tenderlove
30
1.3k
Making Projects Easy
brettharned
120
6.5k
The Art of Delivering Value - GDevCon NA Keynote
reverentgeek
16
1.8k
Making the Leap to Tech Lead
cromwellryan
135
9.6k
Git: the NoSQL Database
bkeepers
PRO
432
66k
Reflections from 52 weeks, 52 projects
jeffersonlam
355
21k
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
• , , =
• • •
• • • •
• •
• • •
• • •
• • • •
•
• • • •
• • •
• • • •
• • •
• • • •
• • • • • • • •
• • • • • • • • •