Intelligence Science and Big Data Engineering. Visual Data Engineering

Name: Intelligence Science and Big Data Engineering. Visual Data Engineering | 9th International Conference, IScIDE 2019, Nanjing, China, October 17-20, 2019, Proceedings, Part I
Brand: Springer
Price: 85.59 EUR
Availability: OnlineOnly

9th International Conference, IScIDE 2019, Nanjing, China, October 17-20, 2019, Proceedings, Part I

Zhen Cui Jinshan Pan Shanshan Zhang Liang Xiao Jian Yang(Editor)

Springer (Publisher)

Published on 28. November 2019

XIX, 577 pages

E-Book

PDF with digital watermarking

System requirements

978-3-030-36189-1 (ISBN)

€85.59incl. 7% vat

System requirements

for PDF with digital watermarking

E-Book Single Licence

Available for download

Description

More details

Other editions

Content

Intro
Preface
Organization
Contents - Part I
Contents - Part II
Deep IA-BI and Five Actions in Circling
1 Deep Bidirectional Intelligence
2 Layered, Topology-Preserved, and Modularly Developing
3 Deep IA-BI Cognition and Image Thinking: From Hubel-Wiesel vs Chen to One Combined Scheme
4 Deep IA-BI Cognition and Abstract Thinking: Searching, Optimising, and Reasoning
5 Summary and Remarks on Split-Brain, IA-BI, and A5
References
Adaptive Online Learning for Video Object Segmentation
1 Introduction
2 The AOL-VOS Method
2.1 Model Foundation
2.2 Adaptive Online Learning for VOS
2.3 Confidence Pattern
3 Experiments
3.1 Datasets
3.2 Results and Comparisons
3.3 Algorithm Analysis
4 Conclusion
References
Proposal-Aware Visual Saliency Detection with Semantic Attention
1 Introduction
2 Proposed Algorithm
2.1 Generate Object Proposals
2.2 Semantic Attention Model
3 Experimental Evaluation
3.1 Datasets
3.2 Implementation Details and Evaluation Criteria
3.3 Performance Comparison
4 Conclusion
References
Constrainted Subspace Low-Rank Representation with Spatial-Spectral Total Variation for Hyperspectral Image Restoration
Abstract
1 Introduction
2 Background
2.1 SLRR Framework
2.2 SSTV Regularization
3 Proposed Method
3.1 CSLRR Framework
3.2 Constrainted Subspace Low-Rank Representation with Spatial-Spectral Total Variation
3.3 Optimization Procedure by IALM
4 Experiments
4.1 Simulated Data Experiments
4.2 Real Data Experiments
5 Conclusion
References
Memory Network-Based Quality Normalization of Magnetic Resonance Images for Brain Segmentation
Abstract
1 Introduction
2 Data
3 Method
3.1 MemNet
3.2 Brain MR Image Segmentation
4 Experiments and Results
4.1 Experimental Settings
4.2 Visual Quality of Normalization Images
4.3 Improvement of Segmentation Accuracy
4.4 Comparative Evaluation
4.5 Complexity
5 Conclusion
Acknowledgement
References
Egomotion Estimation Under Planar Motion with an RGB-D Camera
1 Introduction
2 Ego-Motion Estimation
2.1 The Plane-Plus-Point Algorithm
2.2 The 2-Point Homography-Based Algorithm
3 Experiments
3.1 Test with Synthetic Data
3.2 Performance with Real Data
4 Conclusion
References
Sparse-Temporal Segment Network for Action Recognition
Abstract
1 Introduction
2 The Two-Stream Convolutional Neural Network
3 Sparse-Temporal Segment Network
3.1 Extraction of Sparse Features
3.2 Architecture of Sparse-Temporal Segment Network
4 Experiments
4.1 Datasets
4.2 Different Modality Data Training Network
4.3 Different Input Data
5 Conclusion
Acknowledgment
References
SliceNet: Mask Guided Efficient Feature Augmentation for Attention-Aware Person Re-Identification
1 Introduction
2 Related Work
3 Methodology
3.1 Architecture of SliceNet
3.2 Self-alignment Attention Module
3.3 Loss Function
3.4 Implementation Details
4 Experiments
4.1 Datasets and Evaluation Metrics
4.2 Comparison with State-of-the Arts
4.3 Discussions
5 Conclusion
References
Smoother Soft-NMS for Overlapping Object Detection in X-Ray Images
Abstract
1 Introduction
2 Related Work
3 Approach
3.1 Complex Sample Generation
3.2 Smoother Soft NMS
4 Dataset and Experiment
4.1 Dataset
4.2 Experiment
5 Conclusion
References
Structure-Preserving Guided Image Filtering
1 Introduction
2 Proposed Model and Optimization
2.1 Proposed Model
2.2 Optimization
3 Discussions
3.1 Edge-Preserving Filter
3.2 Gradient-Preserving Filter
3.3 Iterative Filter Kernel
3.4 Filtering Using Color Guidance Image
3.5 Limitations
4 Applications and Experimental Results
4.1 Flash/No-Flash Image Restoration
4.2 Image Dehazing
4.3 Detail Enhancement and HDR Compression
4.4 Image Matting
5 Conclusion
References
Deep Blind Image Inpainting
1 Introduction
2 Related Work
3 Proposed Algorithm
3.1 Network Architecture
3.2 Pre-extracted Gradient Prior
3.3 Loss Function
4 Experimental Results
4.1 Datasets
4.2 Network Training
4.3 Comparisons with State-of-the-Art Methods
5 Analysis and Discussion
5.1 Effect of Feature Extraction Network
5.2 Effect of Pre-extracted Gradient
5.3 Effect of the Loss Function
5.4 Convergence Property
5.5 Additional Baselines
6 Conclusion
References
Robust Object Tracking Based on Multi-granularity Sparse Representation
Abstract
1 Introduction
2 Related Works
3 The Proposed Tracking Framework
3.1 Motion Model
3.2 Observation Model
3.3 Model Update
4 Experiment
4.1 Validation of Our Parameter
4.2 Quantitative Evaluation
4.3 Qualitative Evaluation
5 Conclusion
Acknowledgements
References
A Bypass-Based U-Net for Medical Image Segmentation
Abstract
1 Introduction
2 Bypass-Based U-Net
3 Experimental Evaluation
3.1 Datasets
3.2 Image Pre-processing
3.3 Experimental Setup
3.4 Performance Measures
4 Experimental Results and Discussion
5 Conclusion and Future Works
Acknowledgments
References
Real-Time Visual Object Tracking Based on Reinforcement Learning with Twin Delayed Deep Deterministic Algorithm
1 Introduction
2 Related Work
3 Proposed Method
3.1 Problem Description
3.2 Overview of Our Tracking Framework
3.3 Offline Training
3.4 Online Tracking
3.5 Implementation Details
4 Experiment
4.1 Parameter Setting
4.2 Datasets
4.3 Metrics
4.4 Methods for Comparison
5 Conclusion
References
Efficiently Handling Scale Variation for Pedestrian Detection
1 Introduction
2 Related Work
2.1 Pedestrian Detection with CNNs
2.2 Deep Multi-scale Pedestrian Detection
3 Our Method
3.1 Overview of Proposed Framework
3.2 Multi-stream RCNNs
3.3 Auxiliary Supervision
4 Experiments
4.1 Datasets
4.2 Evaluation Metrics
4.3 Implementation Details
4.4 Comparisons with State-of-the-Art Methods
4.5 Ablation Study
5 Conclusion
References
Leukocyte Segmentation via End-to-End Learning of Deep Convolutional Neural Networks
Abstract
1 Introduction
2 Proposed Method
2.1 Context-Aware Feature Encoder
2.2 Feature Refinement Module
2.3 Feature Decoder
2.4 Loss Function
3 Results
3.1 Dataset and Evaluation Methods
3.2 Implement Details
4 Results and Analysis
4.1 Qualitative Results
4.2 Quantitative Results
5 Conclusion
Acknowledgements
References
Coupled Squeeze-and-Excitation Blocks Based CNN for Image Compression
Abstract
1 Introduction
2 Related Work
3 Coupled SE Learning Framework
3.1 Convolutional Encoder and Decoder
3.2 Coupled SEblocks
3.3 Quantization
3.4 Model Formulation
4 Experiments and Results
4.1 Parameter Description
4.2 Results Evaluation
5 Conclusion
References
Soft Transferring and Progressive Learning for Human Action Recognition
Abstract
1 Introduction
2 Related Works
3 Proposed Method
3.1 Modified DenseNet Structure
3.2 Multi-stage Supervision
3.3 Implementation
4 Experiment Results
4.1 Structure Correlations
4.2 Loss Function Efficacy
4.3 Validation Comparison
5 Conclusions
Acknowledgements
References
Face Sketch Synthesis Based on Adaptive Similarity Regularization
Abstract
1 Introduction
1.1 Prior Works
1.2 Motivation and Contributions
2 Face Sketch Synthesis Based on Adaptive Similarity Regularization
2.1 Adaptive Regularization by Local Similarity
2.2 Adaptive Regularization by Nonlocal Similarity
3 Experimental Results and Analysis
3.1 Parameter Analysis
3.2 Face Sketch Synthesis
4 Conclusions
Acknowledgments
References
Three-Dimensional Coronary Artery Centerline Extraction and Cross Sectional Lumen Quantification from CT Angiography Images
1 Introduction
2 Methodologies
2.1 Centerline Extraction
2.2 Lumen Measurement
3 Results
4 Conclusion
References
A Robust Facial Landmark Detector with Mixed Loss
1 Introduction
2 Related Work
2.1 Pose Variation
2.2 Regression Model
2.3 Loss Function
3 Methodology
3.1 Data Augmentation
3.2 Network and Mixed Loss
3.3 Heatmap to Point Regression
4 Experimental Results
4.1 Datasets
4.2 Experimental Settings
4.3 Results
4.4 Results on AFLW
5 Conclusion
References
Object Guided Beam Steering Algorithm for Optical Phased Array (OPA) LIDAR
1 Introduction
2 Beam Steering
2.1 System Overview
2.2 OPA LIDAR Simulator
2.3 Initial Object Detection
2.4 Beam Allocation
3 Point Cloud Segmentation of OPA LIDAR
4 Experiment
5 Conclusion
References
Channel Max Pooling for Image Classification
1 Introduction
2 The Channel Max Pooling
3 Experiment Results and Discussions
3.1 Experiment Setup
3.2 Evaluation on the Cifar-100 Dataset
3.3 Evaluation on the Care-196 Dataset
3.4 Experiment on the Bilinear Structure
3.5 Discussions
4 Conclusion
References
A Multi-resolution Coarse-to-Fine Segmentation Framework with Active Learning in 3D Brain MRI
Abstract
1 Introduction
2 Related Work
2.1 U-Net like Models in Medical Image Segmentation
2.2 Coarse-to-Fine Segmentation Networks
2.3 Active Learning in Deep Learning
3 Materials and Methods
3.1 Dataset
3.2 Coarse-to-Fine Framework
3.3 Active Learning for Coarse-to-Fine Annotation
4 Experimental Results
4.1 Training
4.2 Segmentation Results
4.3 Comparison with Different Active Learning Strategies
5 Conclusion
Acknowledgements
References
Deep 3D Facial Landmark Detection on Position Maps
1 Introduction
2 Related Work
3 Methodology
3.1 Pose Normalization
3.2 Position Map Generation
3.3 3D Face Landmark Localization Network
4 Experiment and Results
4.1 Databases and Evaluation Criteria
4.2 Comparison to State-of-the-art Methods
4.3 The Role of Regression Network
5 Conclusion
References
Joint Object Detection and Depth Estimation in Multiplexed Image
1 Introduction
2 Related Work
3 Disparity Detector Framework
3.1 The Characteristic of Multiplexed Image
3.2 Disparity Detector
4 Implementation Details
5 Experiments
5.1 Dataset Preparation
5.2 Performance of Object Detection
5.3 Performance of Depth Estimation
6 Conclusion
References
Weakly-Supervised Semantic Segmentation with Mean Teacher Learning
1 Introduction
2 Related Work
2.1 Image-Level Processing
2.2 Seeded Region Growing
3 Proposed Method
3.1 Architecture of the Proposed WSSS-MT
3.2 Mean Teacher Learning
3.3 Seeding and Boundary Loss
4 Experiments
4.1 Experimental Setup
4.2 Compare with Other Methods
4.3 Comparison to the Role of CRF
4.4 Dynamic Observation over Epoch
4.5 Qualitative Results
5 Conclusion
References
APAC-Net: Unsupervised Learning of Depth and Ego-Motion from Monocular Video
1 Introduction
2 APAC-Net
2.1 Depth-Subnet
2.2 Pose-subnet
2.3 Loss Function
3 Experiment
3.1 Depth Estimation
3.2 Pose Estimation
4 Conclusion
References
Robust Image Recovery via Mask Matrix
Abstract
1 Introduction
2 Related Work
3 Masked Robust Principal Component Analysis
3.1 Motivation and Model
3.2 Algorithm
4 Performance Evaluation
4.1 Face Image Recovery Under Lighting Changes
4.2 Face Image Recovery Under Pixel Occlusions
4.3 Background Modeling from Surveillance Video
5 Conclusion
Acknowledgments
References
Multiple Objects Tracking Based Vehicle Speed Analysis with Gaussian Filter from Drone Video
Abstract
1 Introduction
2 Related Work
3 Our Proposed System
3.1 YOLO Detection System
3.2 Kalman Filter
3.3 Gaussian Filter
4 Experiments
4.1 Implementation Details
5 Conclusion
References
A Novel Small Vehicle Detection Method Based on UAV Using Scale Adaptive Gradient Adjustment
Abstract
1 Introduction
2 Related Work
3 Dataset
4 Gradient Adjustment Based on Scale
4.1 Scale Imbalance
4.2 Approaches to Suppress the Problem
4.3 The Relationship Between IOU Change and Scale
4.4 The Disadvantage of YOLOv3
4.5 Our Approach
5 Experiments
6 Conclusion
References
A Level Set Method for Natural Image Segmentation by Texture and High Order Edge-Detector
Abstract
1 Introduction
2 Proposed Method
2.1 The Proposed Texture Term
2.2 The High Order Edge-Detector Term
2.3 Level Set Formulation and Numerical Implementation
3 Experiments
3.1 Texture Feature Comparison
3.2 Segmentation Result Visual Comparison
3.3 Quantitative Comparison
4 Conclusion
Acknowledgements
References
An Attention Bi-box Regression Network for Traffic Light Detection
1 Introduction
2 Related Work
2.1 Traditional Traffic Light Detection
2.2 Traffic Light Detection Based on Deep Learning
3 Approach
3.1 Basic Principle
3.2 Bi-box Regression for Traffic Light
3.3 Attention for Small Object
4 Experiments
4.1 Datasets
4.2 Metrics
4.3 Comparison to State-of-the-Art General Object Detection Algorithm
4.4 Influence of Threshold
4.5 Validity Analysis
5 Conclusion
References
MGD: Mask Guided De-occlusion Framework for Occluded Person Re-identification
1 Introduction
2 Related Work
3 Approach
3.1 Framework
3.2 Coarse-to-Fine Mask Generation Module
3.3 Person De-occlusion Module
4 Experiment
4.1 Datasets
4.2 Experiment Settings
4.3 Comparison with the State-of-the-Art
4.4 Ablation Study
4.5 Mask Evaluation
4.6 Visualization
5 Conclusion
References
Multi-scale Residual Dense Block for Video Super-Resolution
Abstract
1 Introduction
2 Related Work
2.1 Single-Image Super-Resolution
2.2 Video Super-Resolution
3 Proposed Method
3.1 Multi-scale Residual Dense Block (MSRDB)
3.2 Motion Compensation Module
3.3 Multi-scale Residual Dense Network (MSRDN)
3.4 Loss Function
4 Experiments
4.1 Datasets
4.2 Implementation Details
4.3 Experimental Analyses
5 Conclusions
References
Visual Saliency Guided Deep Fabric Defect Classification
Abstract
1 Introduction
2 Related Work
2.1 Regular Method
2.2 Visual Saliency
2.3 Convolutional Neural Network
3 Method
3.1 Overall Architecture
3.2 Visual Saliency Map
3.3 Loss Function
4 Experiments
4.1 Implementation Details
4.2 Data Set Description
4.3 Experimental Setting
4.4 Experiments and Analysis
5 Discussion
6 Conclusion
References
Locality and Sparsity Preserving Embedding Convolutional Neural Network for Image Classification
1 Introduction
2 Related Work
2.1 CNN
2.2 Graph Construction Methods
3 Locality and Sparsity Preserving Embedding CNN
3.1 Motivation to Introduce Manifold Regularization into CNN
3.2 The Framework and Manifold Regularization
3.3 Deep Network Training
4 Experimental Results
4.1 Evaluation Datasets
4.2 Analysis of Experimental Results of CIFAR-10
4.3 Analysis of Experimental Results of CIFAR-100
5 Conclusion
References
Person Re-identification Using Group Constraint
1 Introduction
2 Related Work
2.1 Traditional Re-id Methods
2.2 Pedestrian Detection
3 Framework of Person Re-identification in Groups
3.1 Feature Extraction
3.2 Group Retrieval Correlation
3.3 Fusion of Naive Features and GRC
4 Experiments
4.1 Group-Associated Person Re-id Dataset
4.2 Experimental Settings
4.3 Results
5 Conclusion
References
A Hierarchical Student's t-Distributions Based Unsupervised SAR Image Segmentation Method
Abstract
1 Introduction
2 Preliminaries
2.1 Student's t-Distribution
2.2 Finite Mixture Model
3 Weighted Mean Template Using Structure Tensor
4 FMM Using Hierarchical Student's T-Distributions
5 Parametric Learning Algorithm
6 Experimental Results
6.1 Synthetic Images
6.2 SAR Sea-Ice Images
7 Conclusions
References
Multi-branch Semantic GAN for Infrared Image Generation from Optical Image
1 Introduction
2 Related Work
3 Method
3.1 Multi-generator Training
3.2 Objective
4 Implementation
5 Experiment
5.1 Dataset
5.2 Experimental Results of Scene Classification
5.3 Experimental Results of Image Transformation
6 Conclusion
References
Semantic Segmentation for Prohibited Items in Baggage Inspection
1 Introduction
2 Related Work
3 Method
3.1 Framework
3.2 Channel Module
3.3 Spatial Module
4 Experiments
4.1 Datasets
4.2 Contrast Experiment
4.3 Ablation Experiment
5 Conclusion
References
Sparse Unmixing for Hyperspectral Image with Nonlocal Low-Rank Prior
Abstract
1 Introduction
2 Related Work
2.1 Linear Spectral Unmixing
2.2 Sparse Unmixing
3 Proposed Algorithm
3.1 Nonlocal Self-similarity
3.2 Proposed Model
4 Experimental Results
5 Conclusions
References
Saliency Optimization Integrated Robust Background Detection with Global Ranking
1 Introduction
2 Global Ranking via Absorbed Time in Markov Chain
3 Robust Background Measure
4 Model Construction and Saliency Solution
5 Experiments
5.1 Parameter Setting
5.2 Evaluation Standard
5.3 Global and Robust Cue Validation
5.4 Comparison with Different Methods
6 Conclusion
References
Improvement of Residual Attention Network for Image Classification
1 Introduction
2 Related Work
3 Residual Attention Network
3.1 Improved Network Structure
3.2 Upsampling Method
4 Experiment
4.1 CIFAR-10 and Analysis
4.2 Image Classification on CIFAR-100
5 Conclusion
References
Nuclei Perception Network for Pathology Image Analysis
1 Introduction
2 Related Work
2.1 Nuclei Segmentation in Digital Pathology
2.2 Object Detectors
2.3 Instance Segmentation
3 Methods
3.1 Point Representation of the Nucleus
3.2 Center Point Based Classification Branch
3.3 Instance Segmenation for Nucleus
4 Experiment
4.1 Dataset and Metric
4.2 Results
5 Conclusions
References
A k-Dense-UNet for Biomedical Image Segmentation
1 Introduction
2 Related Work
2.1 Deep Learning Methods for EM Image Segmentation
2.2 DenseNet Architecture
2.3 kU-Net Structure
3 Methods
3.1 Overview of the Proposed Network
3.2 Dense-UNet Architecture
3.3 k-Dense-UNet Formation
4 Experiments and Results
4.1 Dataset and Evaluation Metrics
4.2 Experiments on Loss Function
4.3 Experiments on Differenet Backbones
4.4 Ablation Study
5 Conclusion
References
Gated Fusion of Discriminant Featuresfor Caricature Recognition
1 Introduction
2 Proposed
2.1 Preprocessing
2.2 Network Architecture
2.3 Gated Fusion Unit
2.4 A-Softmax Loss Function
2.5 Implementation
3 Experiments
3.1 Settings
3.2 Ablation Study
3.3 Comparison with State-of-the-art
4 Conclusion
References
Author Index

System requirements

Save as PDF Copy link into clipboard

Schweitzer Fachinformationen

Intelligence Science and Big Data Engineering. Visual Data Engineering

Description

More details

Other editions

Additional editions

Content

System requirements