Submit Search
Upload
Pr057 mask rcnn
•
8 likes
•
4,392 views
T
Taeoh Kim
Follow
Tensorflow Korea 논문읽기 모임 PR12의 57번째 발표는 Instance Segmentation Framework인 Mask R-CNN 입니다
Read less
Read more
Engineering
Report
Share
Report
Share
1 of 73
Download now
Download to read offline
Recommended
Mask R-CNN
Mask R-CNN
Chanuk Lim
Tutorial on Object Detection (Faster R-CNN)
Tutorial on Object Detection (Faster R-CNN)
Hwa Pyung Kim
Yol ov2
Yol ov2
Bang Tsui Liou
Yolo v2 ai_tech_20190421
Yolo v2 ai_tech_20190421
穗碧 陳
Faster R-CNN: Towards real-time object detection with region proposal network...
Faster R-CNN: Towards real-time object detection with region proposal network...
Universitat Politècnica de Catalunya
You only look once
You only look once
Gin Kyeng Lee
Introduction of Faster R-CNN
Introduction of Faster R-CNN
Simossyi Funabashi
Object detection and Instance Segmentation
Object detection and Instance Segmentation
Hichem Felouat
Recommended
Mask R-CNN
Mask R-CNN
Chanuk Lim
Tutorial on Object Detection (Faster R-CNN)
Tutorial on Object Detection (Faster R-CNN)
Hwa Pyung Kim
Yol ov2
Yol ov2
Bang Tsui Liou
Yolo v2 ai_tech_20190421
Yolo v2 ai_tech_20190421
穗碧 陳
Faster R-CNN: Towards real-time object detection with region proposal network...
Faster R-CNN: Towards real-time object detection with region proposal network...
Universitat Politècnica de Catalunya
You only look once
You only look once
Gin Kyeng Lee
Introduction of Faster R-CNN
Introduction of Faster R-CNN
Simossyi Funabashi
Object detection and Instance Segmentation
Object detection and Instance Segmentation
Hichem Felouat
Mask-RCNN for Instance Segmentation
Mask-RCNN for Instance Segmentation
Dat Nguyen
Yolo
Yolo
Bang Tsui Liou
입문 Visual SLAM 14강 - 2장 Introduction to slam
입문 Visual SLAM 14강 - 2장 Introduction to slam
jdo
Yolov5
Yolov5
Hochschule Bonn-Rhein-Sieg
You only look once (YOLO) : unified real time object detection
You only look once (YOLO) : unified real time object detection
Entrepreneur / Startup
PR-207: YOLOv3: An Incremental Improvement
PR-207: YOLOv3: An Incremental Improvement
Jinwon Lee
A Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi Kerola
Preferred Networks
Resnet.pptx
Resnet.pptx
YanhuaSi
Real-time object detection coz YOLO!
Real-time object detection coz YOLO!
J On The Beach
Feature Pyramid Network, FPN
Feature Pyramid Network, FPN
Institute of Agricultural Machinery, NARO
Deep learning for object detection
Deep learning for object detection
Wenjing Chen
Faster R-CNN - PR012
Faster R-CNN - PR012
Jinwon Lee
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
岳華 杜
Recent Progress on Object Detection_20170331
Recent Progress on Object Detection_20170331
Jihong Kang
A brief introduction to recent segmentation methods
A brief introduction to recent segmentation methods
Shunta Saito
Anchor free object detection by deep learning
Anchor free object detection by deep learning
Yu Huang
Yolov3
Yolov3
SHREY MOHAN
Anatomy of YOLO - v1
Anatomy of YOLO - v1
Jihoon Song
論文紹介: Fast R-CNN&Faster R-CNN
論文紹介: Fast R-CNN&Faster R-CNN
Takashi Abe
You Only Look Once: Unified, Real-Time Object Detection
You Only Look Once: Unified, Real-Time Object Detection
DADAJONJURAKUZIEV
Pr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentation
Taeoh Kim
Image-to-Image Translation
Image-to-Image Translation
Junho Kim
More Related Content
What's hot
Mask-RCNN for Instance Segmentation
Mask-RCNN for Instance Segmentation
Dat Nguyen
Yolo
Yolo
Bang Tsui Liou
입문 Visual SLAM 14강 - 2장 Introduction to slam
입문 Visual SLAM 14강 - 2장 Introduction to slam
jdo
Yolov5
Yolov5
Hochschule Bonn-Rhein-Sieg
You only look once (YOLO) : unified real time object detection
You only look once (YOLO) : unified real time object detection
Entrepreneur / Startup
PR-207: YOLOv3: An Incremental Improvement
PR-207: YOLOv3: An Incremental Improvement
Jinwon Lee
A Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi Kerola
Preferred Networks
Resnet.pptx
Resnet.pptx
YanhuaSi
Real-time object detection coz YOLO!
Real-time object detection coz YOLO!
J On The Beach
Feature Pyramid Network, FPN
Feature Pyramid Network, FPN
Institute of Agricultural Machinery, NARO
Deep learning for object detection
Deep learning for object detection
Wenjing Chen
Faster R-CNN - PR012
Faster R-CNN - PR012
Jinwon Lee
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
岳華 杜
Recent Progress on Object Detection_20170331
Recent Progress on Object Detection_20170331
Jihong Kang
A brief introduction to recent segmentation methods
A brief introduction to recent segmentation methods
Shunta Saito
Anchor free object detection by deep learning
Anchor free object detection by deep learning
Yu Huang
Yolov3
Yolov3
SHREY MOHAN
Anatomy of YOLO - v1
Anatomy of YOLO - v1
Jihoon Song
論文紹介: Fast R-CNN&Faster R-CNN
論文紹介: Fast R-CNN&Faster R-CNN
Takashi Abe
You Only Look Once: Unified, Real-Time Object Detection
You Only Look Once: Unified, Real-Time Object Detection
DADAJONJURAKUZIEV
What's hot
(20)
Mask-RCNN for Instance Segmentation
Mask-RCNN for Instance Segmentation
Yolo
Yolo
입문 Visual SLAM 14강 - 2장 Introduction to slam
입문 Visual SLAM 14강 - 2장 Introduction to slam
Yolov5
Yolov5
You only look once (YOLO) : unified real time object detection
You only look once (YOLO) : unified real time object detection
PR-207: YOLOv3: An Incremental Improvement
PR-207: YOLOv3: An Incremental Improvement
A Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi Kerola
Resnet.pptx
Resnet.pptx
Real-time object detection coz YOLO!
Real-time object detection coz YOLO!
Feature Pyramid Network, FPN
Feature Pyramid Network, FPN
Deep learning for object detection
Deep learning for object detection
Faster R-CNN - PR012
Faster R-CNN - PR012
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Recent Progress on Object Detection_20170331
Recent Progress on Object Detection_20170331
A brief introduction to recent segmentation methods
A brief introduction to recent segmentation methods
Anchor free object detection by deep learning
Anchor free object detection by deep learning
Yolov3
Yolov3
Anatomy of YOLO - v1
Anatomy of YOLO - v1
論文紹介: Fast R-CNN&Faster R-CNN
論文紹介: Fast R-CNN&Faster R-CNN
You Only Look Once: Unified, Real-Time Object Detection
You Only Look Once: Unified, Real-Time Object Detection
Similar to Pr057 mask rcnn
Pr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentation
Taeoh Kim
Image-to-Image Translation
Image-to-Image Translation
Junho Kim
Deep Learning for New User Interactions (Gestures, Speech and Emotions)
Deep Learning for New User Interactions (Gestures, Speech and Emotions)
Olivia Klose
On-the-fly Visual Category Search in Web-scale Image Collections
On-the-fly Visual Category Search in Web-scale Image Collections
Ken Chatfield
Lec11 object-re-id
Lec11 object-re-id
United States Air Force Academy
Ilsvrc2015 deep residual_learning_kaiminghe
Ilsvrc2015 deep residual_learning_kaiminghe
pramod naik
[第34回 WBA若手の会勉強会] Microsoft AI platform
[第34回 WBA若手の会勉強会] Microsoft AI platform
Naoki (Neo) SATO
ECCV2010: feature learning for image classification, part 4
ECCV2010: feature learning for image classification, part 4
zukun
Discovering Your AI Super Powers - Tips and Tricks to Jumpstart your AI Projects
Discovering Your AI Super Powers - Tips and Tricks to Jumpstart your AI Projects
Wee Hyong Tok
Class Weighted Convolutional Features for Image Retrieval
Class Weighted Convolutional Features for Image Retrieval
Universitat Politècnica de Catalunya
Deep Learning for Computer Vision: Segmentation (UPC 2016)
Deep Learning for Computer Vision: Segmentation (UPC 2016)
Universitat Politècnica de Catalunya
Auro tripathy - Localizing with CNNs
Auro tripathy - Localizing with CNNs
Auro Tripathy
Deep Learning for Computer Vision: Object Detection (UPC 2016)
Deep Learning for Computer Vision: Object Detection (UPC 2016)
Universitat Politècnica de Catalunya
D3L4-objects.pdf
D3L4-objects.pdf
ssusere945ae
Pixel RNN to Pixel CNN++
Pixel RNN to Pixel CNN++
Dongheon Lee
Object Detection - Míriam Bellver - UPC Barcelona 2018
Object Detection - Míriam Bellver - UPC Barcelona 2018
Universitat Politècnica de Catalunya
Object Segmentation (D2L7 Insight@DCU Machine Learning Workshop 2017)
Object Segmentation (D2L7 Insight@DCU Machine Learning Workshop 2017)
Universitat Politècnica de Catalunya
The impact of visual saliency prediction in image classification
The impact of visual saliency prediction in image classification
Universitat Politècnica de Catalunya
Windows to reality getting the most out of direct3 d 10 graphics in your games
Windows to reality getting the most out of direct3 d 10 graphics in your games
changehee lee
Interpretability of Convolutional Neural Networks - Eva Mohedano - UPC Barcel...
Interpretability of Convolutional Neural Networks - Eva Mohedano - UPC Barcel...
Universitat Politècnica de Catalunya
Similar to Pr057 mask rcnn
(20)
Pr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentation
Image-to-Image Translation
Image-to-Image Translation
Deep Learning for New User Interactions (Gestures, Speech and Emotions)
Deep Learning for New User Interactions (Gestures, Speech and Emotions)
On-the-fly Visual Category Search in Web-scale Image Collections
On-the-fly Visual Category Search in Web-scale Image Collections
Lec11 object-re-id
Lec11 object-re-id
Ilsvrc2015 deep residual_learning_kaiminghe
Ilsvrc2015 deep residual_learning_kaiminghe
[第34回 WBA若手の会勉強会] Microsoft AI platform
[第34回 WBA若手の会勉強会] Microsoft AI platform
ECCV2010: feature learning for image classification, part 4
ECCV2010: feature learning for image classification, part 4
Discovering Your AI Super Powers - Tips and Tricks to Jumpstart your AI Projects
Discovering Your AI Super Powers - Tips and Tricks to Jumpstart your AI Projects
Class Weighted Convolutional Features for Image Retrieval
Class Weighted Convolutional Features for Image Retrieval
Deep Learning for Computer Vision: Segmentation (UPC 2016)
Deep Learning for Computer Vision: Segmentation (UPC 2016)
Auro tripathy - Localizing with CNNs
Auro tripathy - Localizing with CNNs
Deep Learning for Computer Vision: Object Detection (UPC 2016)
Deep Learning for Computer Vision: Object Detection (UPC 2016)
D3L4-objects.pdf
D3L4-objects.pdf
Pixel RNN to Pixel CNN++
Pixel RNN to Pixel CNN++
Object Detection - Míriam Bellver - UPC Barcelona 2018
Object Detection - Míriam Bellver - UPC Barcelona 2018
Object Segmentation (D2L7 Insight@DCU Machine Learning Workshop 2017)
Object Segmentation (D2L7 Insight@DCU Machine Learning Workshop 2017)
The impact of visual saliency prediction in image classification
The impact of visual saliency prediction in image classification
Windows to reality getting the most out of direct3 d 10 graphics in your games
Windows to reality getting the most out of direct3 d 10 graphics in your games
Interpretability of Convolutional Neural Networks - Eva Mohedano - UPC Barcel...
Interpretability of Convolutional Neural Networks - Eva Mohedano - UPC Barcel...
More from Taeoh Kim
CNN Attention Networks
CNN Attention Networks
Taeoh Kim
PR 127: FaceNet
PR 127: FaceNet
Taeoh Kim
PR 113: The Perception Distortion Tradeoff
PR 113: The Perception Distortion Tradeoff
Taeoh Kim
PR 103: t-SNE
PR 103: t-SNE
Taeoh Kim
Pr083 Non-local Neural Networks
Pr083 Non-local Neural Networks
Taeoh Kim
Pr072 deep compression
Pr072 deep compression
Taeoh Kim
More from Taeoh Kim
(6)
CNN Attention Networks
CNN Attention Networks
PR 127: FaceNet
PR 127: FaceNet
PR 113: The Perception Distortion Tradeoff
PR 113: The Perception Distortion Tradeoff
PR 103: t-SNE
PR 103: t-SNE
Pr083 Non-local Neural Networks
Pr083 Non-local Neural Networks
Pr072 deep compression
Pr072 deep compression
Recently uploaded
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Low Rate Call Girls In Saket, Delhi NCR
Electronically Controlled suspensions system .pdf
Electronically Controlled suspensions system .pdf
me23b1001
Concrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptx
KartikeyaDwivedi3
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024
Mark Billinghurst
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
Suhani Kapoor
Application of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptx
959SahilShah
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
Asst.prof M.Gokilavani
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
DeepakSakkari2
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
9953056974 Low Rate Call Girls In Saket, Delhi NCR
Churning of Butter, Factors affecting .
Churning of Butter, Factors affecting .
Satyam Kumar
power system scada applications and uses
power system scada applications and uses
DevarapalliHaritha
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentation
GDSCAESB
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptx
britheesh05
Internship report on mechanical engineering
Internship report on mechanical engineering
malavadedarshan25
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
RajaP95
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
Tsuyoshi Horigome
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
9953056974 Low Rate Call Girls In Saket, Delhi NCR
Call Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call Girls
ssuser7cb4ff
Introduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptx
vipinkmenon1
Design and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdf
Tagore Institute of Engineering And Technology
Recently uploaded
(20)
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
Electronically Controlled suspensions system .pdf
Electronically Controlled suspensions system .pdf
Concrete Mix Design - IS 10262-2019 - .pptx
Concrete Mix Design - IS 10262-2019 - .pptx
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
Application of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptx
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
Churning of Butter, Factors affecting .
Churning of Butter, Factors affecting .
power system scada applications and uses
power system scada applications and uses
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentation
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptx
Internship report on mechanical engineering
Internship report on mechanical engineering
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call Girls
Introduction to Microprocesso programming and interfacing.pptx
Introduction to Microprocesso programming and interfacing.pptx
Design and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdf
Pr057 mask rcnn
1.
Yonsei University MVP Lab.
2.
3.
Bbox Regression Classification RoI from Selective Search RoI Pooling FixedSizeRepresentation
4.
Bbox Regression Classification RoI Pooling FixedSizeRepresentation Bbox Regression Objectness RPN Region Proposal Network
5.
32x32x3 Conv1 Pool1 16x16x64 Conv2 Pool2 8x8x128 Conv3 Pool3 4x4x256 Conv4 Pool4 2x2x512 Conv5 Pool5 1x1x512 1x1x512 Conv 1x1 Heatmap x32
Upsample Softmax Remove Pooling 1x1 Conv for Heatmap Output
6.
7.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
8.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
9.
Sheep Dog Human Sheep Sheep Sheep
Sheep
10.
Sheep Dog Human
11.
Dog Human Sheep Sheep Sheep Sheep Sheep
12.
BBox Classification Segmentation Classification
13.
BBox Classification Segmentation Classification Can Separate Cannot Segment
14.
BBox Classification Segmentation Classification Can Separate Cannot Segment Cannot
Separate Can Segment
15.
BBox Classification Segmentation Classification Segmentation in BBox Classification + = Can
Separate Cannot Segment Cannot Separate Can Segment
16.
BBox Classification Segmentation Classification Segmentation in BBox Classification + = Can
Separate Cannot Segment Cannot Separate Can Segment Faster R-CNN FCN
17.
BBox Classification Segmentation Classification Segmentation in BBox Classification Faster R-CNN
FCN FCN on BBOX ! + = + = Can Separate Cannot Segment Cannot Separate Can Segment
18.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
FCN • Pixel-level Classification •
Per Pixel Softmax (Multinomial) • Multi Instance
29.
FCN • Pixel-level Classification •
Per Pixel Softmax (Multinomial) • Multi Instance Faster R-CNN • Classification • Instance Level RoI
30.
FCN • Pixel-level Classification •
Per Pixel Softmax (Multinomial) • Multi Instance Faster R-CNN • Classification • Instance Level RoI
31.
FCN • Pixel-level Classification •
Per Pixel Softmax Sigmoid (Binary) • Multi Instance Faster R-CNN • Classification • Instance Level RoI
32.
FCN • Pixel-level Classification •
Per Pixel Softmax Sigmoid (Binary) • Multi Instance Faster R-CNN • Classification • Instance Level RoI
33.
DB BBox + Class
+ Mask 𝐿 = 𝐿𝑐𝑙𝑠 + 𝐿 𝑏𝑜𝑥 + 𝐿 𝑚𝑎𝑠𝑘 𝐿𝑐𝑙𝑠: Softmax Cross Entropy 𝐿 𝑏𝑜𝑥: Regression 𝐿 𝑚𝑎𝑠𝑘: Binary Cross Entropy
34.
Training Phase 𝐿 𝑚𝑎𝑠𝑘
= 𝐿𝑐1 + 𝐿𝑐2 + ⋯+ 𝐿𝑐𝑘 𝐿 𝑚𝑎𝑠𝑘 = 𝐿𝑐3 if) GT Class is 3
35.
Training Phase 𝐿 𝑚𝑎𝑠𝑘
= 𝐿𝑐1 + 𝐿𝑐2 + ⋯+ 𝐿𝑐𝑘 𝐿 𝑚𝑎𝑠𝑘 = 𝐿𝑐3 if) GT Class is 3 Mask Branch Only Learns How to Mask independent of Class
36.
Test Phase Predicts Human
Mask Predicts Car Mask Predicts Horse Mask Predicts ...
37.
Test Phase Predicts Human
Mask Predicts Car Mask Predicts Horse Mask Predicts ... Winner Takes All
38.
39.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
40.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
41.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017 FasterR-CNN,S.Ren,NIPS2015
42.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017 Deconv 2x2 str2 Deconv 2x2
str2
43.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017 3x3
Conv 4 Layer
44.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017 1x1 Conv 1x1
Conv
45.
46.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
47.
Bbox Regression Classification RoI Pooling FixedSizeRepresentation Pooled Feature 7x7
48.
RoI Pooling (Fast
R-CNN) • Input: Each RoI • Output: 7x7 Pooled Feature RoI Align (Mask R-CNN) • Input: Each RoI • Output: 7x7 Pooled Feature
49.
RoI Pooling (Fast
R-CNN) • Input: Each RoI • Output: 7x7 Pooled Feature RoI Align (Mask R-CNN) • Input: Each RoI • Output: 7x7 Pooled Feature
50.
Feature Map RoI Note: Region Proposal
Network RoI Prediction = Floating Point Representation
51.
Feature Map RoI
52.
Feature Map RoI
53.
Feature Map RoI Max Pooling
54.
Feature Map RoI Max Pooling
55.
Feature Map RoI
56.
Feature Map RoI
57.
Feature Map RoI 2x2 Subcells
for Precision
58.
= 0.15 +
0.25 + 0.25 + 0.35 RoI
59.
Feature Map RoI 2x2 Subcell
Max Pooling
60.
Bbox Regression Classification RoI Align Bbox Regression Objectness RPN Binary Mask
61.
Bbox Regression Classification RoI Align Bbox Regression Objectness RPN Binary Mask Paste
Back
62.
SlidefromMaskR-CNNTutorial, K.He.ICCV2017
63.
64.
• Faster R-CNN
+ ResNet Deep ResidualLearning for Image Recognition, K He, 2016 CVPR • Faster R-CNN + FPN Feature Pyramid Networks for Object Detection, T.Y.Lin 2017 CVPR
65.
• Faster R-CNN
+ ResNet Deep ResidualLearning for Image Recognition, K He, 2016 CVPR
66.
• Faster R-CNN
+ FPN Feature Pyramid Networks for Object Detection, T.Y.Lin 2017 CVPR
67.
68.
Faster R-CNN +
Binary Mask Prediction + FCN + RoIAlign
69.
Faster R-CNN +
Binary Mask Prediction + FCN + RoIAlign
70.
Detection Performance Improvement
71.
72.
73.
Q&A?
Download now