SlideShare a Scribd company logo
1 of 22
Download to read offline
Detection in Crowded Scenes:
One Proposal, Multiple Predictions
Xuangeng Chu, Anlin Zheng, Xiangyu Zhang, Jian Sun
Peking University, MEGVII Technology
CVPR 2020
2021.11.21
딥러닝논문읽기모임 이미지처리팀
홍은기, 김병현, 김선옥, 안종식, 이찬혁
2
목차
1. Introduction
2. Proposed Approach: Multiple Instance Prediction
3. Experiment
4. Conclusion & Discussion
3
Chu et al., 2020, Detection in Crowded Scenes: One Proposal, Multiple Predictions
Introduction – Crowded Object Detection
4
1. Proposed a novel approach: Multiple Instance Prediction
2. Proposed a novel loss: EMD loss
3. Proposed a novel NMS: Set NMS
4. Achieved SOTA on CrowdHuman Dataset
Contribution
5
- Shao et al., 2018, CrowdHuman: A Benchmark or Detection Human in a Crowd
- https://www.crowdhuman.org/
• train/val/test: 15,000 / 4,370 / 5,000
• 470K human instances
CrowdHuman Dataset
6
Chu et al., 2020, Detection in Crowded Scenes: One Proposal, Multiple Predictions
• State-of-the-art models on COCO or VOC perform poorly on CrowdHuman dataset
1) Highly overlapped instances are likely to have very similar features
2) Heavily overlapped instances are likely to be mistakenly suppressed by NMS
Fundamental difficulties in crowded object detection
7
https://towardsdatascience.com/non-maximum-suppression-nms-93ce178e177c
NMS (Non-Maximum Suppression)
8
• For each proposal box, rather than predicting a single instance, propose a set of instances
Solution – multiple instance prediction
(a) Each proposal box predicts a single instance
(intrinsically difficult!). After NMS, only one
prediction survives.
(b) Set NMS removes duplicates from different
proposals while keeping duplicates in a proposal.
single prediction
paradigm
multiple instance
prediction
9
• Step 1: assign a proposal box to ground-truths
Solution – multiple instance prediction
proposal b1
g1
g2
g3
10
• Step 2: make K predictions from one proposal box
Solution – multiple instance prediction
proposal b1
g1
g2
g3
p1
p2
p3
K = 3
11
• Step 3: assign predictions to ground-truths using Earth Mover’s Distance (EMD)
EMD Loss
p1
P2
P3
g1
g2
g3
background
EMD loss:
g1
g2
g3
p1
p2
p3
K = 3
12
• Step 4: apply Set NMS
Set NMS
Set NMS
13
Set NMS
14
Architecture
15
Q & A
16
Experiments
• Evaluation Metrics
1) Averaged Precision (AP)
2) MR-2 Miss Rate on False Positive Per Image (FPPI) in [10-2, 100])
3) Jaccard Index
• Datasets
1) CrowdHuman
2) CityPersons
3) COCO
• Network Architecture
1) Backbone: ResNet-50 pre-trained on ImageNet
2) Head: FPN with RoIAlign
3) K = 2
17
Main results and ablation study
Performance on CrowdHuman Dataset
18
Comparison with various NMS strategies
Performance on CrowdHuman Dataset
19
Ablation on Number of Heads
Performance on CrowdHuman Dataset
20
Experiments on COCO
21
Conclusion & Discussion
1. Proposed approach is not only effective on crowded scenes, but also generalizes well on
normal data.
2. Proposed approach is compatible with other one-stage & two-stage architectures.
3. A local version of DETR (Carion et al., 2020)?
22
Thank you

More Related Content

What's hot

Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)
Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)
Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)Universitat Politècnica de Catalunya
 
MN-3, MN-Core and HPL - SC21 Green500 BOF
MN-3, MN-Core and HPL - SC21 Green500 BOFMN-3, MN-Core and HPL - SC21 Green500 BOF
MN-3, MN-Core and HPL - SC21 Green500 BOFPreferred Networks
 
Graph Convolutional Neural Networks
Graph Convolutional Neural Networks Graph Convolutional Neural Networks
Graph Convolutional Neural Networks 신동 강
 
Introduction to 3D Computer Vision and Differentiable Rendering
Introduction to 3D Computer Vision and Differentiable RenderingIntroduction to 3D Computer Vision and Differentiable Rendering
Introduction to 3D Computer Vision and Differentiable RenderingPreferred Networks
 
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...Seiya Ito
 
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...Taegyun Jeon
 
FCN-Based 6D Robotic Grasping for Arbitrary Placed Objects
FCN-Based 6D Robotic Grasping for Arbitrary Placed ObjectsFCN-Based 6D Robotic Grasping for Arbitrary Placed Objects
FCN-Based 6D Robotic Grasping for Arbitrary Placed ObjectsKusano Hitoshi
 
TensorFlow Tutorial Part1
TensorFlow Tutorial Part1TensorFlow Tutorial Part1
TensorFlow Tutorial Part1Sungjoon Choi
 
NIPS読み会2013: One-shot learning by inverting a compositional causal process
NIPS読み会2013: One-shot learning by inverting  a compositional causal processNIPS読み会2013: One-shot learning by inverting  a compositional causal process
NIPS読み会2013: One-shot learning by inverting a compositional causal processnozyh
 
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic SegmentationSemantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation岳華 杜
 
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...Jumlesha Shaik
 
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...Alex Conway
 
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)Universitat Politècnica de Catalunya
 
Deep learning for molecules, introduction to chainer chemistry
Deep learning for molecules, introduction to chainer chemistryDeep learning for molecules, introduction to chainer chemistry
Deep learning for molecules, introduction to chainer chemistryKenta Oono
 
Beyond data and model parallelism for deep neural networks
Beyond data and model parallelism for deep neural networksBeyond data and model parallelism for deep neural networks
Beyond data and model parallelism for deep neural networksJunKudo2
 
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...Joonhyung Lee
 
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Lalit Jain
 
Semantic segmentation with Convolutional Neural Network Approaches
Semantic segmentation with Convolutional Neural Network ApproachesSemantic segmentation with Convolutional Neural Network Approaches
Semantic segmentation with Convolutional Neural Network ApproachesFellowship at Vodafone FutureLab
 
Pr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentationPr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentationTaeoh Kim
 

What's hot (20)

Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)
Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)
Optimization for Deep Networks (D2L1 2017 UPC Deep Learning for Computer Vision)
 
MN-3, MN-Core and HPL - SC21 Green500 BOF
MN-3, MN-Core and HPL - SC21 Green500 BOFMN-3, MN-Core and HPL - SC21 Green500 BOF
MN-3, MN-Core and HPL - SC21 Green500 BOF
 
Graph Convolutional Neural Networks
Graph Convolutional Neural Networks Graph Convolutional Neural Networks
Graph Convolutional Neural Networks
 
Introduction to 3D Computer Vision and Differentiable Rendering
Introduction to 3D Computer Vision and Differentiable RenderingIntroduction to 3D Computer Vision and Differentiable Rendering
Introduction to 3D Computer Vision and Differentiable Rendering
 
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...
[3D勉強会@関東] Deep Reinforcement Learning of Volume-guided Progressive View Inpa...
 
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...
[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...
 
FCN-Based 6D Robotic Grasping for Arbitrary Placed Objects
FCN-Based 6D Robotic Grasping for Arbitrary Placed ObjectsFCN-Based 6D Robotic Grasping for Arbitrary Placed Objects
FCN-Based 6D Robotic Grasping for Arbitrary Placed Objects
 
TensorFlow Tutorial Part1
TensorFlow Tutorial Part1TensorFlow Tutorial Part1
TensorFlow Tutorial Part1
 
NIPS読み会2013: One-shot learning by inverting a compositional causal process
NIPS読み会2013: One-shot learning by inverting  a compositional causal processNIPS読み会2013: One-shot learning by inverting  a compositional causal process
NIPS読み会2013: One-shot learning by inverting a compositional causal process
 
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic SegmentationSemantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
Semantic Segmentation - Fully Convolutional Networks for Semantic Segmentation
 
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...
Ml srhwt-machine-learning-based-superlative-rapid-haar-wavelet-transformation...
 
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
 
Deep Generative Models - Kevin McGuinness - UPC Barcelona 2018
Deep Generative Models - Kevin McGuinness - UPC Barcelona 2018Deep Generative Models - Kevin McGuinness - UPC Barcelona 2018
Deep Generative Models - Kevin McGuinness - UPC Barcelona 2018
 
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)
Optimizing Deep Networks (D1L6 Insight@DCU Machine Learning Workshop 2017)
 
Deep learning for molecules, introduction to chainer chemistry
Deep learning for molecules, introduction to chainer chemistryDeep learning for molecules, introduction to chainer chemistry
Deep learning for molecules, introduction to chainer chemistry
 
Beyond data and model parallelism for deep neural networks
Beyond data and model parallelism for deep neural networksBeyond data and model parallelism for deep neural networks
Beyond data and model parallelism for deep neural networks
 
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...
DeepLab V3+: Encoder-Decoder with Atrous Separable Convolution for Semantic I...
 
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow) Object classification using CNN & VGG16 Model (Keras and Tensorflow)
Object classification using CNN & VGG16 Model (Keras and Tensorflow)
 
Semantic segmentation with Convolutional Neural Network Approaches
Semantic segmentation with Convolutional Neural Network ApproachesSemantic segmentation with Convolutional Neural Network Approaches
Semantic segmentation with Convolutional Neural Network Approaches
 
Pr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentationPr045 deep lab_semantic_segmentation
Pr045 deep lab_semantic_segmentation
 

Similar to 211121 detection in crowded scenes one proposal, multiple predictions

A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...
A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...
A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...Angie Miller
 
最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に - 最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に - Hiroshi Fukui
 
Object Detection and Tracking using Statistical and Stochastic Techniques
Object Detection and Tracking using Statistical and Stochastic TechniquesObject Detection and Tracking using Statistical and Stochastic Techniques
Object Detection and Tracking using Statistical and Stochastic TechniquesVasuhiSamydurai1
 
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)Towards Accurate Multi-person Pose Estimation in the Wild (My summery)
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)Abdulrahman Kerim
 
myashar_research_2016
myashar_research_2016myashar_research_2016
myashar_research_2016Mark Yashar
 
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...grssieee
 
A framework for outlier detection in
A framework for outlier detection inA framework for outlier detection in
A framework for outlier detection inijfcstjournal
 
PPT - Deep and Confident Prediction For Time Series at Uber
PPT - Deep and Confident Prediction For Time Series at UberPPT - Deep and Confident Prediction For Time Series at Uber
PPT - Deep and Confident Prediction For Time Series at UberJisang Yoon
 
Human detection in hours of
Human detection in hours ofHuman detection in hours of
Human detection in hours ofijistjournal
 
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...Tarik Reza Toha
 
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...Wanjin Yu
 
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...multimediaeval
 
"The effect of angles and distance on image-based three-dimensional reconstru...
"The effect of angles and distance on image-based three-dimensional reconstru..."The effect of angles and distance on image-based three-dimensional reconstru...
"The effect of angles and distance on image-based three-dimensional reconstru...TRUSS ITN
 
Ieee 2016 nss mic poster N30-21
Ieee 2016 nss mic poster N30-21Ieee 2016 nss mic poster N30-21
Ieee 2016 nss mic poster N30-21Dae Woon Kim
 
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...Edge AI and Vision Alliance
 
PR157: Best of both worlds: human-machine collaboration for object annotation
PR157: Best of both worlds: human-machine collaboration for object annotationPR157: Best of both worlds: human-machine collaboration for object annotation
PR157: Best of both worlds: human-machine collaboration for object annotationjaewon lee
 
"An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ..."An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ...butest
 
Time-delayed collective flow diffusion models for inferring latent people flo...
Time-delayed collective flow diffusion models for inferring latent people flo...Time-delayed collective flow diffusion models for inferring latent people flo...
Time-delayed collective flow diffusion models for inferring latent people flo...Shun Kojima
 

Similar to 211121 detection in crowded scenes one proposal, multiple predictions (20)

A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...
A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...
A Gans-Based Deep Learning Framework For Automatic Subsurface Object Recognit...
 
最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に - 最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に -
 
Object Detection and Tracking using Statistical and Stochastic Techniques
Object Detection and Tracking using Statistical and Stochastic TechniquesObject Detection and Tracking using Statistical and Stochastic Techniques
Object Detection and Tracking using Statistical and Stochastic Techniques
 
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)Towards Accurate Multi-person Pose Estimation in the Wild (My summery)
Towards Accurate Multi-person Pose Estimation in the Wild (My summery)
 
myashar_research_2016
myashar_research_2016myashar_research_2016
myashar_research_2016
 
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...
VERIFICATION_&_VALIDATION_OF_A_SEMANTIC_IMAGE_TAGGING_FRAMEWORK_VIA_GENERATIO...
 
A framework for outlier detection in
A framework for outlier detection inA framework for outlier detection in
A framework for outlier detection in
 
PPT - Deep and Confident Prediction For Time Series at Uber
PPT - Deep and Confident Prediction For Time Series at UberPPT - Deep and Confident Prediction For Time Series at Uber
PPT - Deep and Confident Prediction For Time Series at Uber
 
Human detection in hours of
Human detection in hours ofHuman detection in hours of
Human detection in hours of
 
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...
Beyond Counting: Comparisons of Density Maps for Crowd Analysis Tasks—Countin...
 
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...
Human Behavior Understanding: From Human-Oriented Analysis to Action Recognit...
 
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...
Classification of Strokes in Table Tennis with a Three Stream Spatio-Temporal...
 
"The effect of angles and distance on image-based three-dimensional reconstru...
"The effect of angles and distance on image-based three-dimensional reconstru..."The effect of angles and distance on image-based three-dimensional reconstru...
"The effect of angles and distance on image-based three-dimensional reconstru...
 
20210226 esa-science-coffee-v2.0
20210226 esa-science-coffee-v2.020210226 esa-science-coffee-v2.0
20210226 esa-science-coffee-v2.0
 
Ieee 2016 nss mic poster N30-21
Ieee 2016 nss mic poster N30-21Ieee 2016 nss mic poster N30-21
Ieee 2016 nss mic poster N30-21
 
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...
“Understanding DNN-Based Object Detectors,” a Presentation from Au-Zone Techn...
 
PR157: Best of both worlds: human-machine collaboration for object annotation
PR157: Best of both worlds: human-machine collaboration for object annotationPR157: Best of both worlds: human-machine collaboration for object annotation
PR157: Best of both worlds: human-machine collaboration for object annotation
 
"An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ..."An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ...
 
Time-delayed collective flow diffusion models for inferring latent people flo...
Time-delayed collective flow diffusion models for inferring latent people flo...Time-delayed collective flow diffusion models for inferring latent people flo...
Time-delayed collective flow diffusion models for inferring latent people flo...
 
Binary Analysis - Luxembourg
Binary Analysis - LuxembourgBinary Analysis - Luxembourg
Binary Analysis - Luxembourg
 

More from taeseon ryu

OpineSum Entailment-based self-training for abstractive opinion summarization...
OpineSum Entailment-based self-training for abstractive opinion summarization...OpineSum Entailment-based self-training for abstractive opinion summarization...
OpineSum Entailment-based self-training for abstractive opinion summarization...taeseon ryu
 
3D Gaussian Splatting
3D Gaussian Splatting3D Gaussian Splatting
3D Gaussian Splattingtaeseon ryu
 
Hyperbolic Image Embedding.pptx
Hyperbolic  Image Embedding.pptxHyperbolic  Image Embedding.pptx
Hyperbolic Image Embedding.pptxtaeseon ryu
 
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정taeseon ryu
 
LLaMA Open and Efficient Foundation Language Models - 230528.pdf
LLaMA Open and Efficient Foundation Language Models - 230528.pdfLLaMA Open and Efficient Foundation Language Models - 230528.pdf
LLaMA Open and Efficient Foundation Language Models - 230528.pdftaeseon ryu
 
Dataset Distillation by Matching Training Trajectories
Dataset Distillation by Matching Training Trajectories Dataset Distillation by Matching Training Trajectories
Dataset Distillation by Matching Training Trajectories taeseon ryu
 
Packed Levitated Marker for Entity and Relation Extraction
Packed Levitated Marker for Entity and Relation ExtractionPacked Levitated Marker for Entity and Relation Extraction
Packed Levitated Marker for Entity and Relation Extractiontaeseon ryu
 
MOReL: Model-Based Offline Reinforcement Learning
MOReL: Model-Based Offline Reinforcement LearningMOReL: Model-Based Offline Reinforcement Learning
MOReL: Model-Based Offline Reinforcement Learningtaeseon ryu
 
Scaling Instruction-Finetuned Language Models
Scaling Instruction-Finetuned Language ModelsScaling Instruction-Finetuned Language Models
Scaling Instruction-Finetuned Language Modelstaeseon ryu
 
Visual prompt tuning
Visual prompt tuningVisual prompt tuning
Visual prompt tuningtaeseon ryu
 
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdf
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdfvariBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdf
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdftaeseon ryu
 
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdf
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdfReinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdf
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdftaeseon ryu
 
The Forward-Forward Algorithm
The Forward-Forward AlgorithmThe Forward-Forward Algorithm
The Forward-Forward Algorithmtaeseon ryu
 
Towards Robust and Reproducible Active Learning using Neural Networks
Towards Robust and Reproducible Active Learning using Neural NetworksTowards Robust and Reproducible Active Learning using Neural Networks
Towards Robust and Reproducible Active Learning using Neural Networkstaeseon ryu
 
BRIO: Bringing Order to Abstractive Summarization
BRIO: Bringing Order to Abstractive SummarizationBRIO: Bringing Order to Abstractive Summarization
BRIO: Bringing Order to Abstractive Summarizationtaeseon ryu
 

More from taeseon ryu (20)

VoxelNet
VoxelNetVoxelNet
VoxelNet
 
OpineSum Entailment-based self-training for abstractive opinion summarization...
OpineSum Entailment-based self-training for abstractive opinion summarization...OpineSum Entailment-based self-training for abstractive opinion summarization...
OpineSum Entailment-based self-training for abstractive opinion summarization...
 
3D Gaussian Splatting
3D Gaussian Splatting3D Gaussian Splatting
3D Gaussian Splatting
 
JetsonTX2 Python
 JetsonTX2 Python  JetsonTX2 Python
JetsonTX2 Python
 
Hyperbolic Image Embedding.pptx
Hyperbolic  Image Embedding.pptxHyperbolic  Image Embedding.pptx
Hyperbolic Image Embedding.pptx
 
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정
MCSE_Multimodal Contrastive Learning of Sentence Embeddings_변현정
 
LLaMA Open and Efficient Foundation Language Models - 230528.pdf
LLaMA Open and Efficient Foundation Language Models - 230528.pdfLLaMA Open and Efficient Foundation Language Models - 230528.pdf
LLaMA Open and Efficient Foundation Language Models - 230528.pdf
 
YOLO V6
YOLO V6YOLO V6
YOLO V6
 
Dataset Distillation by Matching Training Trajectories
Dataset Distillation by Matching Training Trajectories Dataset Distillation by Matching Training Trajectories
Dataset Distillation by Matching Training Trajectories
 
RL_UpsideDown
RL_UpsideDownRL_UpsideDown
RL_UpsideDown
 
Packed Levitated Marker for Entity and Relation Extraction
Packed Levitated Marker for Entity and Relation ExtractionPacked Levitated Marker for Entity and Relation Extraction
Packed Levitated Marker for Entity and Relation Extraction
 
MOReL: Model-Based Offline Reinforcement Learning
MOReL: Model-Based Offline Reinforcement LearningMOReL: Model-Based Offline Reinforcement Learning
MOReL: Model-Based Offline Reinforcement Learning
 
Scaling Instruction-Finetuned Language Models
Scaling Instruction-Finetuned Language ModelsScaling Instruction-Finetuned Language Models
Scaling Instruction-Finetuned Language Models
 
Visual prompt tuning
Visual prompt tuningVisual prompt tuning
Visual prompt tuning
 
mPLUG
mPLUGmPLUG
mPLUG
 
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdf
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdfvariBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdf
variBAD, A Very Good Method for Bayes-Adaptive Deep RL via Meta-Learning.pdf
 
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdf
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdfReinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdf
Reinforced Genetic Algorithm Learning For Optimizing Computation Graphs.pdf
 
The Forward-Forward Algorithm
The Forward-Forward AlgorithmThe Forward-Forward Algorithm
The Forward-Forward Algorithm
 
Towards Robust and Reproducible Active Learning using Neural Networks
Towards Robust and Reproducible Active Learning using Neural NetworksTowards Robust and Reproducible Active Learning using Neural Networks
Towards Robust and Reproducible Active Learning using Neural Networks
 
BRIO: Bringing Order to Abstractive Summarization
BRIO: Bringing Order to Abstractive SummarizationBRIO: Bringing Order to Abstractive Summarization
BRIO: Bringing Order to Abstractive Summarization
 

Recently uploaded

Real-Time AI Streaming - AI Max Princeton
Real-Time AI  Streaming - AI Max PrincetonReal-Time AI  Streaming - AI Max Princeton
Real-Time AI Streaming - AI Max PrincetonTimothy Spann
 
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档208367051
 
Vision, Mission, Goals and Objectives ppt..pptx
Vision, Mission, Goals and Objectives ppt..pptxVision, Mission, Goals and Objectives ppt..pptx
Vision, Mission, Goals and Objectives ppt..pptxellehsormae
 
RABBIT: A CLI tool for identifying bots based on their GitHub events.
RABBIT: A CLI tool for identifying bots based on their GitHub events.RABBIT: A CLI tool for identifying bots based on their GitHub events.
RABBIT: A CLI tool for identifying bots based on their GitHub events.natarajan8993
 
RadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfRadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfgstagge
 
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...Biometric Authentication: The Evolution, Applications, Benefits and Challenge...
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...GQ Research
 
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝DelhiRS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhijennyeacort
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...Amil Baba Dawood bangali
 
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...ssuserf63bd7
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxMike Bennett
 
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改yuu sss
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryJeremy Anderson
 
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Boston Institute of Analytics
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]📊 Markus Baersch
 
Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Seán Kennedy
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFAAndrei Kaleshka
 
Heart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectHeart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectBoston Institute of Analytics
 
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesConf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesTimothy Spann
 
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一F sss
 

Recently uploaded (20)

Real-Time AI Streaming - AI Max Princeton
Real-Time AI  Streaming - AI Max PrincetonReal-Time AI  Streaming - AI Max Princeton
Real-Time AI Streaming - AI Max Princeton
 
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
 
Vision, Mission, Goals and Objectives ppt..pptx
Vision, Mission, Goals and Objectives ppt..pptxVision, Mission, Goals and Objectives ppt..pptx
Vision, Mission, Goals and Objectives ppt..pptx
 
RABBIT: A CLI tool for identifying bots based on their GitHub events.
RABBIT: A CLI tool for identifying bots based on their GitHub events.RABBIT: A CLI tool for identifying bots based on their GitHub events.
RABBIT: A CLI tool for identifying bots based on their GitHub events.
 
RadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfRadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdf
 
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...Biometric Authentication: The Evolution, Applications, Benefits and Challenge...
Biometric Authentication: The Evolution, Applications, Benefits and Challenge...
 
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝DelhiRS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
 
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...
Statistics, Data Analysis, and Decision Modeling, 5th edition by James R. Eva...
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptx
 
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data Story
 
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
 
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]
 
Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFA
 
Heart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectHeart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis Project
 
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesConf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
 
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
 

211121 detection in crowded scenes one proposal, multiple predictions

  • 1. Detection in Crowded Scenes: One Proposal, Multiple Predictions Xuangeng Chu, Anlin Zheng, Xiangyu Zhang, Jian Sun Peking University, MEGVII Technology CVPR 2020 2021.11.21 딥러닝논문읽기모임 이미지처리팀 홍은기, 김병현, 김선옥, 안종식, 이찬혁
  • 2. 2 목차 1. Introduction 2. Proposed Approach: Multiple Instance Prediction 3. Experiment 4. Conclusion & Discussion
  • 3. 3 Chu et al., 2020, Detection in Crowded Scenes: One Proposal, Multiple Predictions Introduction – Crowded Object Detection
  • 4. 4 1. Proposed a novel approach: Multiple Instance Prediction 2. Proposed a novel loss: EMD loss 3. Proposed a novel NMS: Set NMS 4. Achieved SOTA on CrowdHuman Dataset Contribution
  • 5. 5 - Shao et al., 2018, CrowdHuman: A Benchmark or Detection Human in a Crowd - https://www.crowdhuman.org/ • train/val/test: 15,000 / 4,370 / 5,000 • 470K human instances CrowdHuman Dataset
  • 6. 6 Chu et al., 2020, Detection in Crowded Scenes: One Proposal, Multiple Predictions • State-of-the-art models on COCO or VOC perform poorly on CrowdHuman dataset 1) Highly overlapped instances are likely to have very similar features 2) Heavily overlapped instances are likely to be mistakenly suppressed by NMS Fundamental difficulties in crowded object detection
  • 8. 8 • For each proposal box, rather than predicting a single instance, propose a set of instances Solution – multiple instance prediction (a) Each proposal box predicts a single instance (intrinsically difficult!). After NMS, only one prediction survives. (b) Set NMS removes duplicates from different proposals while keeping duplicates in a proposal. single prediction paradigm multiple instance prediction
  • 9. 9 • Step 1: assign a proposal box to ground-truths Solution – multiple instance prediction proposal b1 g1 g2 g3
  • 10. 10 • Step 2: make K predictions from one proposal box Solution – multiple instance prediction proposal b1 g1 g2 g3 p1 p2 p3 K = 3
  • 11. 11 • Step 3: assign predictions to ground-truths using Earth Mover’s Distance (EMD) EMD Loss p1 P2 P3 g1 g2 g3 background EMD loss: g1 g2 g3 p1 p2 p3 K = 3
  • 12. 12 • Step 4: apply Set NMS Set NMS Set NMS
  • 16. 16 Experiments • Evaluation Metrics 1) Averaged Precision (AP) 2) MR-2 Miss Rate on False Positive Per Image (FPPI) in [10-2, 100]) 3) Jaccard Index • Datasets 1) CrowdHuman 2) CityPersons 3) COCO • Network Architecture 1) Backbone: ResNet-50 pre-trained on ImageNet 2) Head: FPN with RoIAlign 3) K = 2
  • 17. 17 Main results and ablation study Performance on CrowdHuman Dataset
  • 18. 18 Comparison with various NMS strategies Performance on CrowdHuman Dataset
  • 19. 19 Ablation on Number of Heads Performance on CrowdHuman Dataset
  • 21. 21 Conclusion & Discussion 1. Proposed approach is not only effective on crowded scenes, but also generalizes well on normal data. 2. Proposed approach is compatible with other one-stage & two-stage architectures. 3. A local version of DETR (Carion et al., 2020)?