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
Paper-Survey: Objects as Points
Search
fam_taro
April 19, 2019
Science
0
2.3k
Paper-Survey: Objects as Points
fam_taro
April 19, 2019
Tweet
Share
More Decks by fam_taro
See All by fam_taro
NeRFの概要と 派生系についてのふんわり紹介
fam_taro
3
3.9k
実践 PyTorch Lightning (2019/11/30 分析コンペLT会 #1)
fam_taro
3
4.4k
Paper:ShapeMask
fam_taro
0
57
Summary: Objects as Points
fam_taro
0
3.1k
Tensorコアを使った PyTorch の高速化について
fam_taro
4
3.8k
Sequence to Sequence Learning with Neural Networks
fam_taro
1
1k
Other Decks in Science
See All in Science
はじめての「相関と因果とエビデンス」入門:“動機づけられた推論” に抗うために
takehikoihayashi
17
6.8k
作業領域内の障害物を回避可能なバイナリマニピュレータの設計 / Design of binary manipulator avoiding obstacles in workspace
konakalab
0
160
Spectral Sparsification of Hypergraphs
tasusu
0
170
The Incredible Machine: Developer Productivity and the Impact of AI
tomzimmermann
0
390
事業会社における 機械学習・推薦システム技術の活用事例と必要な能力 / ml-recsys-in-layerx-wantedly-2024
yuya4
3
230
Science of Scienceおよび科学計量学に関する研究論文の俯瞰可視化_ポスター版
hayataka88
0
130
Celebrate UTIG: Staff and Student Awards 2024
utig
0
460
第61回コンピュータビジョン勉強会「BioCLIP: A Vision Foundation Model for the Tree of Life」
x_ttyszk
1
1.5k
ほたるのひかり/RayTracingCamp10
kugimasa
0
210
Mechanistic Interpretability の紹介
sohtakahashi
0
350
Direct Preference Optimization
zchenry
0
280
はじめてのバックドア基準:あるいは、重回帰分析の偏回帰係数を因果効果の推定値として解釈してよいのか問題
takehikoihayashi
2
740
Featured
See All Featured
Optimising Largest Contentful Paint
csswizardry
33
2.9k
What's new in Ruby 2.0
geeforr
343
31k
Producing Creativity
orderedlist
PRO
341
39k
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
mongodb
42
9.2k
XXLCSS - How to scale CSS and keep your sanity
sugarenia
246
1.3M
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
10
720
Reflections from 52 weeks, 52 projects
jeffersonlam
346
20k
Visualization
eitanlees
145
15k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
6
420
Put a Button on it: Removing Barriers to Going Fast.
kastner
59
3.5k
Facilitating Awesome Meetings
lara
50
6.1k
Done Done
chrislema
181
16k
Transcript
จLT: Objects as Points h"ps:/ /arxiv.org/abs/1904.07850 2019/04/19 ౻ຊ༟հ 1
࣍ • ஶऀใ • ֓ཁ • ͜Ε·ͰͷϞσϧͱͷҧ͍ • ਫ਼ •
ͦͷଞײ 2
ஶऀใ • Xingyi Zhou(UT Aus1n) • Dequan Wang(UC Berkeley) •
Philipp Krähenbühl(UT Aus1n) 3
ಛ • ମݕग़Ϟσϧ • ༗໊ͳྫ: SSD, YOLOv3, Re.naNet, M2Det... •
ݕग़ͷΈͳΒͣ࢟ɾdepthɾ͖ɾ3d size ʹద༻͍ͯ͠Δ • backbone ͱͯ͠ DLA(deep layers aggrega.on) Hourglass(CornerNet Ͱ ༻) Λ༻ 4
ಛ • bounding box ΛΘͣʹݕग़Λߦ͏Ϟσϧ(keypointਪఆ) • bounding box ༻ͷ grid
ͷΘΓʹ͕ࡉ͔͍ heatmap(H, W Λ4Ͱׂͬͨఔ ͷͷ) Λग़ྗ • heatmap ͕ߴ͍ॴ() Λମͷத৺ͱਪఆ • த৺ͱͳΔॴͷ feature ͔Βମͷେ͖͞ɾࢄԽޡࠩΛਪఆ • ࢄԽޡࠩ = heatmap ʹͨ͠ࡍͷޡࠩ • େ͖͞ʹ͍ͭͯ scale ͍ͯ͠ͳ͍(ͦͷ··ͷ) 5
ಛ • ༧ଌϘοΫε = heatmap ͷ࠲ඪ + ༧ଌϘοΫεαΠζ + ༧ଌࢄԽޡࠩ
• ֶशʹ͏ heatmap ͷ 1ମʹ͖ͭ 1ͭͷΈ • SSD ͷΑ͏ʹ IoU ͷॏͳΓ۩߹Ͱ background ͔൱͔Λ͚ͳ͍ • ෳ box ग़͞ͳ͍͜ͱΛલఏͱ͍ͯ͠Δ • ಉ͡ΫϥεͰॏͳͬͯ͠·͏߹͕͋Δ͕શମͷ 0.1 % ະຬͰ RCNN(2% ະ ຬ) ΑΓখ͍͞ 6
Πϝʔδਤ 7
͜Ε·ͰͷϞσϧͱͷҧ͍ • Object detec*on with implicit anchors(SSD, YOLO, Re*naNet )ͱͷҧ͍
• CenterNetശͷॏͳΓͰͳ͘ҐஔͷΈʹج͍ͮͯʮΞϯΧʔʯΛׂ • લܠͱഎܠͷྨʹؔ͢Δखಈͷ͖͍͠ͳ͍(IoU 0.5 > ͱ͔) • ମຖʹϙδςΟϒͳΞϯΧʔ1͚ͭͩͳͷͰ NMS Λඞཁͱ͠ͳ͍ • We simply extract local peaks in the keypoint heatmap • keypoint heatmap ͔ΒϩʔΧϧϐʔΫΛநग़͢Δ͚ͩͰྑ͍ 8
͜Ε·ͰͷϞσϧͱͷҧ͍ • Object detec*on with implicit anchors(SSD, YOLO, Re*naNet )ͱͷҧ͍
• CenterNetΑΓେ͖ͳग़ྗղ૾Λ͏ • mask r-cnn ͱ͔ͱൺֱͯ͠ • output stride of 16 • ͜ΕʹΑΓෳͷΞϯΧʔ͕ෆཁͱͳΔʁʁʁʁ • [1711.08189] An Analysis of Scale Invariance in Object Detec*on - SNIP 9
͜Ε·ͰͷϞσϧͱͷҧ͍ • Object detec*on by keypoint es*ma*on(CornerNet, ExtremeNet )ͱͷҧ͍ •
্ه 2ͭ keypoint ݕग़ޙʹ Έ߹ΘͤΛ grouping ͢Δඞཁ͕͋Δ • ͘ͳͬͯ͠·͏ • CenterNet ඞཁͱ͠ͳ͍ • ͍ʂ 10
ਫ਼ 11
ਫ਼(M2Det ͷ݁ՌΛࢹͰՃͯ͠Έͨ) 12
ͦͷଞײ • Backbone ͱͯ͠ DLA Λ͑ΔͷΛॳΊͯͬͨ • Ή͠Ζ DLA ॳΊͯΓ·ͨ͠
! • NMS ͕ෆཁʹͳΔͷຯʹخ͍͠ • anchor ͕ফ͑Δͷخ͍͠ • খ͍͞ମʹରͯ͠ͲΕ͚ͩରԠͰ͖Δ͔֬ೝ͠ͳ͍ͱ 13