Fully Convolutional Networks for Semantic Segmentation

Abstract

Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves improved segmentation of PASCAL VOC (30% relative improvement to 67.2% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.

to appear in PAMI (accepted May, 2016); journal edition of arXiv:1411.4038

References in corpus (19)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
• Multi-Scale Context Aggregation by Dilated Convolutions
• Going Deeper with Convolutions
• ParseNet: Looking Wider to See Better
• Learning Deconvolution Network for Semantic Segmentation
• Rich feature hierarchies for accurate object detection and semantic segmentation
• Convolutional Feature Masking for Joint Object and Stuff Segmentation
• Descriptor Matching with Convolutional Neural Networks: a Comparison to SIFT
• Simultaneous Detection and Segmentation
• Decoupled Deep Neural Network for Semi-supervised Semantic Segmentation
• Building high-level features using large scale unsupervised learning
• BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation
• Fast Image Scanning with Deep Max-Pooling Convolutional Neural Networks
• Feedforward semantic segmentation with zoom-out features
• Pedestrian Detection with Unsupervised Multi-Stage Feature Learning

Cited by in corpus (396)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs
• ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data
• Densely Connected Convolutional Networks
• Focal Loss for Dense Object Detection
• The Liver Tumor Segmentation Benchmark (LiTS)
• ParseNet: Looking Wider to See Better
• FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
• Learning Deconvolution Network for Semantic Segmentation
• A Deep Learning Approach for Brain Tumor Classification and Segmentation Using a Multiscale Convolutional Neural Network
• Eigen-CAM: Class Activation Map using Principal Components
• Convolutional Feature Masking for Joint Object and Stuff Segmentation
• Asymmetric Tri-training for Unsupervised Domain Adaptation
• Optic Disc and Cup Segmentation Methods for Glaucoma Detection with Modification of U-Net Convolutional Neural Network
• Deformable Convolutional Networks
• Convolutional neural networks with low-rank regularization
• Fast-SCNN: Fast Semantic Segmentation Network
• Generalizing to Unseen Domains via Adversarial Data Augmentation
• Weakly- and Semi-Supervised Learning of a DCNN for Semantic Image Segmentation
• Evolution of Image Segmentation using Deep Convolutional Neural Network: A Survey
• PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation
• Residual Attention Network for Image Classification
• Pyramid Scene Parsing Network
• Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities and Results of the WMH Segmentation Challenge
• FusionNet: A deep fully residual convolutional neural network for image segmentation in connectomics
• LocalViT: Analyzing Locality in Vision Transformers
• End-to-End Learning of Geometry and Context for Deep Stereo Regression
• A Review of Object Detection Models based on Convolutional Neural Network
• Vehicle Detection from 3D Lidar Using Fully Convolutional Network
• ABC-CNN: An Attention Based Convolutional Neural Network for Visual Question Answering
• Unsupervised Visual Representation Learning by Context Prediction
• Fully Convolutional Multi-Class Multiple Instance Learning
• A2D2: Audi Autonomous Driving Dataset
• Fully Connected Deep Structured Networks
• FSS-1000: A 1000-Class Dataset for Few-Shot Segmentation
• Image Captioning with Semantic Attention
• Rotation equivariant vector field networks
• A Survey on Deep Learning-based Architectures for Semantic Segmentation on 2D images
• Wider or Deeper: Revisiting the ResNet Model for Visual Recognition
• Asymmetric Loss Functions and Deep Densely Connected Networks for Highly Imbalanced Medical Image Segmentation: Application to Multiple Sclerosis Lesion Detection
• Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network
• Automatic Liver and Tumor Segmentation of CT and MRI Volumes using Cascaded Fully Convolutional Neural Networks
• Dilated Residual Networks
• Discriminative Localization in CNNs for Weakly-Supervised Segmentation of Pulmonary Nodules
• Deep Learning for Time-Series Analysis
• An Entropy-based Pruning Method for CNN Compression
• CNN-based Segmentation of Medical Imaging Data
• Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Bag of Tricks for Image Classification with Convolutional Neural Networks
• Deep Contrast Learning for Salient Object Detection
• Context Encoding for Semantic Segmentation
• Good Practice in CNN Feature Transfer
• Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecture
• Learning to cluster in order to transfer across domains and tasks
• Variable Rate Image Compression with Recurrent Neural Networks
• Amulet: Aggregating Multi-level Convolutional Features for Salient Object Detection
• Unsupervised Deep Embedding for Clustering Analysis
• Multi-style Generative Network for Real-time Transfer
• Proposal-free Network for Instance-level Object Segmentation
• Deep Learning Face Attributes in the Wild
• Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform
• Deep Reinforcement Learning for Visual Object Tracking in Videos
• What Do We Understand About Convolutional Networks?
• Learning Deep Feature Representations with Domain Guided Dropout for Person Re-identification
• Coin.AI: A Proof-of-Useful-Work Scheme for Blockchain-based Distributed Deep Learning
• Rethinking ImageNet Pre-training
• Fast Online Object Tracking and Segmentation: A Unifying Approach
• Transfer Learning from Deep Features for Remote Sensing and Poverty Mapping
• DSOD: Learning Deeply Supervised Object Detectors from Scratch
• PixelNet: Representation of the pixels, by the pixels, and for the pixels
• MAttNet: Modular Attention Network for Referring Expression Comprehension
• Learning a Discriminative Feature Network for Semantic Segmentation
• Multi-Oriented Text Detection with Fully Convolutional Networks
• Automatic Annotation of Hip Anatomy in Fluoroscopy for Robust and Efficient 2D/3D Registration
• Photographic Image Synthesis with Cascaded Refinement Networks
• Building a Large Scale Dataset for Image Emotion Recognition: The Fine Print and The Benchmark
• Learning Multi-Domain Convolutional Neural Networks for Visual Tracking
• Makeup like a superstar: Deep Localized Makeup Transfer Network
• A PCB Dataset for Defects Detection and Classification
• Learning Background-Aware Correlation Filters for Visual Tracking
• Pyramid Stereo Matching Network
• A Survey of Semantic Segmentation
• PAD-Net: Multi-Tasks Guided Prediction-and-Distillation Network for Simultaneous Depth Estimation and Scene Parsing
• Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks
• Real-Time Well Log Prediction From Drilling Data Using Deep Learning
• CDC: Convolutional-De-Convolutional Networks for Precise Temporal Action Localization in Untrimmed Videos
• Analysis and Optimization of Convolutional Neural Network Architectures
• Optimizing Sparse RFI Prediction using Deep Learning
• Unsupervised Histopathology Image Synthesis
• PixelNet: Towards a General Pixel-level Architecture
• Deep Direct Regression for Multi-Oriented Scene Text Detection
• ShiftCNN: Generalized Low-Precision Architecture for Inference of Convolutional Neural Networks
• Semi and Weakly Supervised Semantic Segmentation Using Generative Adversarial Network
• Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks
• Deep Cocktail Network: Multi-source Unsupervised Domain Adaptation with Category Shift
• Deep Variation-structured Reinforcement Learning for Visual Relationship and Attribute Detection
• Colorization as a Proxy Task for Visual Understanding
• FISHING Net: Future Inference of Semantic Heatmaps In Grids
• Photographic Text-to-Image Synthesis with a Hierarchically-nested Adversarial Network
• Adversarial Dropout Regularization
• A Localisation-Segmentation Approach for Multi-label Annotation of Lumbar Vertebrae using Deep Nets
• Pushing the Boundaries of Boundary Detection using Deep Learning
• Deeply-Recursive Convolutional Network for Image Super-Resolution
• Unsupervised Deep Multi-focus Image Fusion
• SegFlow: Joint Learning for Video Object Segmentation and Optical Flow
• Training Deep Networks with Structured Layers by Matrix Backpropagation
• CANet: Class-Agnostic Segmentation Networks with Iterative Refinement and Attentive Few-Shot Learning
• EvalAI: Towards Better Evaluation Systems for AI Agents
• WordSup: Exploiting Word Annotations for Character based Text Detection
• Deep Learning Convolutional Networks for Multiphoton Microscopy Vasculature Segmentation
• Deep Discriminative Clustering Analysis
• Learning deep structured active contours end-to-end
• Data Distillation: Towards Omni-Supervised Learning
• Deep Interactive Object Selection
• Annotating Object Instances with a Polygon-RNN
• Deep Learning for Semantic Part Segmentation with High-Level Guidance
• Structured Attention Guided Convolutional Neural Fields for Monocular Depth Estimation
• Two-Phase Learning for Weakly Supervised Object Localization
• Accounting for Variance in Machine Learning Benchmarks
• Sparseness Meets Deepness: 3D Human Pose Estimation from Monocular Video
• Weakly Supervised Instance Segmentation using Class Peak Response
• Fully Convolutional Networks for Chip-wise Defect Detection Employing Photoluminescence Images
• Deep learning approach for identification of HII regions during reionization in 21-cm observations
• Deep Watershed Transform for Instance Segmentation
• Semantic Image Segmentation with Task-Specific Edge Detection Using CNNs and a Discriminatively Trained Domain Transform
• SPGNet: Semantic Prediction Guidance for Scene Parsing
• Monocular Object Instance Segmentation and Depth Ordering with CNNs
• Unconstrained Facial Landmark Localization with Backbone-Branches Fully-Convolutional Networks
• Deep Convolutional Neural Networks with Merge-and-Run Mappings
• Temporal Generative Adversarial Nets with Singular Value Clipping
• Deep Layer Aggregation
• Learning Better Features for Face Detection with Feature Fusion and Segmentation Supervision
• Learning in an Uncertain World: Representing Ambiguity Through Multiple Hypotheses
• Exploiting saliency for object segmentation from image level labels
• DeepMVS: Learning Multi-view Stereopsis
• Semantic Object Parsing with Local-Global Long Short-Term Memory
• Learning Analysis-by-Synthesis for 6D Pose Estimation in RGB-D Images
• Object Detection, Tracking, and Motion Segmentation for Object-level Video Segmentation
• Deep Convolutional Features for Image Based Retrieval and Scene Categorization
• Deep Pyramidal Residual Networks
• Class-incremental Learning via Deep Model Consolidation
• Siamese Cascaded Region Proposal Networks for Real-Time Visual Tracking
• Scaling Wide Residual Networks for Panoptic Segmentation
• ZM-Net: Real-time Zero-shot Image Manipulation Network
• SCAN: Structure Correcting Adversarial Network for Organ Segmentation in Chest X-rays
• Weakly Supervised Dense Video Captioning
• Video Propagation Networks
• High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis
• 3D Object Reconstruction from a Single Depth View with Adversarial Learning
• Learning Accurate Low-Bit Deep Neural Networks with Stochastic Quantization
• Joint Sequence Learning and Cross-Modality Convolution for 3D Biomedical Segmentation
• Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentation
• FCN-rLSTM: Deep Spatio-Temporal Neural Networks for Vehicle Counting in City Cameras
• CASENet: Deep Category-Aware Semantic Edge Detection
• Electricity Theft Detection with self-attention
• 3D Fully Convolutional Network for Vehicle Detection in Point Cloud
• Semantic Instance Annotation of Street Scenes by 3D to 2D Label Transfer
• Learning Uncertain Convolutional Features for Accurate Saliency Detection
• InstanceCut: from Edges to Instances with MultiCut
• Natural Language Guided Visual Relationship Detection
• No More Discrimination: Cross City Adaptation of Road Scene Segmenters
• Improving Fully Convolution Network for Semantic Segmentation
• Lung cancer screening with low-dose CT scans using a deep learning approach
• Lattice Long Short-Term Memory for Human Action Recognition
• Weakly-supervised Visual Grounding of Phrases with Linguistic Structures
• Automatic localization and decoding of honeybee markers using deep convolutional neural networks
• Agile Amulet: Real-Time Salient Object Detection with Contextual Attention
• Learning a Discriminative Prior for Blind Image Deblurring
• PPR-FCN: Weakly Supervised Visual Relation Detection via Parallel Pairwise R-FCN
• Knowledge Adaptation for Efficient Semantic Segmentation
• Improved Techniques for Training Adaptive Deep Networks
• Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields
• Microparticle cloud imaging and tracking for data-driven plasma science
• Learning to Segment Instances in Videos with Spatial Propagation Network
• 3D Shape Induction from 2D Views of Multiple Objects
• All about Structure: Adapting Structural Information across Domains for Boosting Semantic Segmentation
• Deep Image Harmonization
• A survey of Object Classification and Detection based on 2D/3D data
• TorontoCity: Seeing the World with a Million Eyes
• Scene Graph Generation via Conditional Random Fields
• Self-supervised Low Light Image Enhancement and Denoising
• ILabel: Interactive Neural Scene Labelling
• A Pixel-Based Framework for Data-Driven Clothing
• MonoCap: Monocular Human Motion Capture using a CNN Coupled with a Geometric Prior
• Learning Digital Camera Pipeline for Extreme Low-Light Imaging
• Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes
• Visualizing and Understanding Deep Texture Representations
• EPINET: A Fully-Convolutional Neural Network Using Epipolar Geometry for Depth from Light Field Images
• Borrowing Treasures from the Wealthy: Deep Transfer Learning through Selective Joint Fine-tuning
• Few-shot Object Detection via Feature Reweighting
• Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervision
• Structured Binary Neural Networks for Accurate Image Classification and Semantic Segmentation
• Semantic Single-Image Dehazing
• Object Detection Free Instance Segmentation With Labeling Transformations
• Instance Embedding Transfer to Unsupervised Video Object Segmentation
• C3: Concentrated-Comprehensive Convolution and its application to semantic segmentation
• Learning-based Defect Recognition for Quasi-Periodic Microscope Images
• Learning to Segment Every Thing
• A New Convolutional Network-in-Network Structure and Its Applications in Skin Detection, Semantic Segmentation, and Artifact Reduction
• Detecting Oriented Text in Natural Images by Linking Segments
• Deep Extreme Cut: From Extreme Points to Object Segmentation
• 3D Human Pose Estimation from a Single Image via Distance Matrix Regression
• Cross Domain Knowledge Transfer for Person Re-identification
• Pixel-wise Segmentation of Street with Neural Networks
• Single-Shot Object Detection with Enriched Semantics
• Understanding Traffic Density from Large-Scale Web Camera Data
• Multisource and Multitemporal Data Fusion in Remote Sensing
• ACE: Adapting to Changing Environments for Semantic Segmentation
• Visual Discovery at Pinterest
• DeepEdge: A Multi-Scale Bifurcated Deep Network for Top-Down Contour Detection
• An Iterative Convolutional Neural Network Algorithm Improves Electron Microscopy Image Segmentation
• Attention Gated Networks: Learning to Leverage Salient Regions in Medical Images
• LCNN: Lookup-based Convolutional Neural Network
• Skin Lesion Segmentation and Classification for ISIC 2018 by Combining Deep CNN and Handcrafted Features
• Pose2Instance: Harnessing Keypoints for Person Instance Segmentation
• TAFE-Net: Task-Aware Feature Embeddings for Low Shot Learning
• Noise2Void - Learning Denoising from Single Noisy Images
• SRN: Side-output Residual Network for Object Symmetry Detection in the Wild
• Label-Driven Reconstruction for Domain Adaptation in Semantic Segmentation
• Reversible Recursive Instance-level Object Segmentation
• Iterative Deep Learning for Network Topology Extraction
• Pointwise Convolutional Neural Networks
• DSAC - Differentiable RANSAC for Camera Localization
• Learning Sparse High Dimensional Filters: Image Filtering, Dense CRFs and Bilateral Neural Networks
• Learning High Dynamic Range from Outdoor Panoramas
• SpherePHD: Applying CNNs on a Spherical PolyHeDron Representation of 360 degree Images
• Deep End2End Voxel2Voxel Prediction
• AI-based Aortic Vessel Tree Segmentation for Cardiovascular Diseases Treatment: Status Quo
• Joint Multi-Person Pose Estimation and Semantic Part Segmentation
• Learning Transferrable Knowledge for Semantic Segmentation with Deep Convolutional Neural Network
• Data-driven Seismic Waveform Inversion: A Study on the Robustness and Generalization
• Semantic Segmentation for Partially Occluded Apple Trees Based on Deep Learning
• AdaCoSeg: Adaptive Shape Co-Segmentation with Group Consistency Loss
• Is Image Super-resolution Helpful for Other Vision Tasks?
• Enhancing Cross-task Black-Box Transferability of Adversarial Examples with Dispersion Reduction
• Bidirectional Attention Network for Monocular Depth Estimation
• Learning a Discriminative Filter Bank within a CNN for Fine-grained Recognition
• OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold
• Self-Supervised Feature Learning by Learning to Spot Artifacts
• Boundary-aware Instance Segmentation
• Toward Streaming Synapse Detection with Compositional ConvNets
• Visual Compiler: Synthesizing a Scene-Specific Pedestrian Detector and Pose Estimator
• Towards Automatic Image Editing: Learning to See another You
• Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training
• Fast Semantic Image Segmentation with High Order Context and Guided Filtering
• Dense Captioning with Joint Inference and Visual Context
• Weakly Supervised Object Localization Using Things and Stuff Transfer
• Super-Resolution with Deep Adaptive Image Resampling
• Face Parsing via Recurrent Propagation
• SketchParse : Towards Rich Descriptions for Poorly Drawn Sketches using Multi-Task Hierarchical Deep Networks
• Deep Quantization: Encoding Convolutional Activations with Deep Generative Model
• Learning High-level Prior with Convolutional Neural Networks for Semantic Segmentation
• ProNet: Learning to Propose Object-specific Boxes for Cascaded Neural Networks
• Deep Structured Scene Parsing by Learning with Image Descriptions
• Deep Regionlets for Object Detection
• Semantic Object Parsing with Graph LSTM
• Weakly and Semi Supervised Human Body Part Parsing via Pose-Guided Knowledge Transfer
• A 3D Coarse-to-Fine Framework for Volumetric Medical Image Segmentation
• Surveillance Video Parsing with Single Frame Supervision
• Domain adaptation for holistic skin detection
• Dense Recurrent Neural Networks for Scene Labeling
• Adaptive Weighting Multi-Field-of-View CNN for Semantic Segmentation in Pathology
• Analysis on DeepLabV3+ Performance for Automatic Steel Defects Detection
• Error Correction for Dense Semantic Image Labeling
• A Unified Framework for Generalizable Style Transfer: Style and Content Separation
• Learning to detect and localize many objects from few examples
• ArbiText: Arbitrary-Oriented Text Detection in Unconstrained Scene
• A Multiscale Patch Based Convolutional Network for Brain Tumor Segmentation
• DeepSkeleton: Skeleton Map for 3D Human Pose Regression
• Auxiliary Learning for Deep Multi-task Learning
• Irregular Convolutional Neural Networks
• Part-Stacked CNN for Fine-Grained Visual Categorization
• Large-Scale 3D Scene Classification With Multi-View Volumetric CNN
• Colorectal Polyp Segmentation by U-Net with Dilation Convolution
• Semi-supervised Skin Detection by Network with Mutual Guidance
• Scene Understanding Networks for Autonomous Driving based on Around View Monitoring System
• Contour Detection Using Cost-Sensitive Convolutional Neural Networks
• Distantly Supervised Road Segmentation
• Attend in groups: a weakly-supervised deep learning framework for learning from web data
• Beyond One Glance: Gated Recurrent Architecture for Hand Segmentation
• Simultaneous Segmentation and Relaxometry for MRI through Multitask Learning
• Semantic Segmentation of Skin Lesions using a Small Data Set
• Greenhouse Segmentation on High-Resolution Optical Satellite Imagery using Deep Learning Techniques
• Super-BPD: Super Boundary-to-Pixel Direction for Fast Image Segmentation
• Dual Encoder Fusion U-Net (DEFU-Net) for Cross-manufacturer Chest X-ray Segmentation
• Multi-domain semantic segmentation with pyramidal fusion
• SeesawFaceNets: sparse and robust face verification model for mobile platform
• Automatic Generation of Dense Non-rigid Optical Flow
• Attention Mechanisms for Object Recognition with Event-Based Cameras
• Zoom and Learn: Generalizing Deep Stereo Matching to Novel Domains
• Detecting Small, Densely Distributed Objects with Filter-Amplifier Networks and Loss Boosting
• Neuron-level Selective Context Aggregation for Scene Segmentation
• Cross-Domain Self-supervised Multi-task Feature Learning using Synthetic Imagery
• SeGAN: Segmenting and Generating the Invisible
• Physically-Based Rendering for Indoor Scene Understanding Using Convolutional Neural Networks
• Learning Motion Patterns in Videos
• On Deep Representation Learning from Noisy Web Images
• Superpixel Convolutional Networks using Bilateral Inceptions
• Generalized Deep Image to Image Regression
• Learning Deep Representations for Scene Labeling with Semantic Context Guided Supervision
• Point and Ask: Incorporating Pointing into Visual Question Answering
• TopoResNet: A hybrid deep learning architecture and its application to skin lesion classification
• Multilevel Context Representation for Improving Object Recognition
• Frustum VoxNet for 3D object detection from RGB-D or Depth images
• CNN based texture synthesize with Semantic segment
• AutoScaler: Scale-Attention Networks for Visual Correspondence
• Not Using the Car to See the Sidewalk: Quantifying and Controlling the Effects of Context in Classification and Segmentation
• Trinity: A No-Code AI platform for complex spatial datasets
• DeepFlux for Skeletons in the Wild
• Investigating Transfer Learning Capabilities of Vision Transformers and CNNs by Fine-Tuning a Single Trainable Block
• Butterfly Effect: Bidirectional Control of Classification Performance by Small Additive Perturbation
• Unsupervised Domain Adaptation using Generative Models and Self-ensembling
• SurfConv: Bridging 3D and 2D Convolution for RGBD Images
• A Systematic Comparison of Deep Learning Architectures in an Autonomous Vehicle
• Unsupervised Learning of Important Objects from First-Person Videos
• Deep Reflectance Maps
• Real-time Driver Drowsiness Detection for Android Application Using Deep Neural Networks Techniques
• Fast and Accurate Online Video Object Segmentation via Tracking Parts
• Defective samples simulation through Neural Style Transfer for automatic surface defect segment
• Semi-supervised Semantic Segmentation with Directional Context-aware Consistency
• Unsupervised learning with sparse space-and-time autoencoders
• End-to-End Saliency Mapping via Probability Distribution Prediction
• Mango Tree Net -- A fully convolutional network for semantic segmentation and individual crown detection of mango trees
• Detect, Replace, Refine: Deep Structured Prediction For Pixel Wise Labeling
• DeMeshNet: Blind Face Inpainting for Deep MeshFace Verification
• Instance-Level Salient Object Segmentation
• IMP: Instance Mask Projection for High Accuracy Semantic Segmentation of Things
• Feature-Fused Context-Encoding Network for Neuroanatomy Segmentation
• Learning Where to Look: Data-Driven Viewpoint Set Selection for 3D Scenes
• Spatially-Adaptive Filter Units for Deep Neural Networks
• Dynamic Video Segmentation Network
• Diverse Sampling for Self-Supervised Learning of Semantic Segmentation
• Explainable Semantic Mapping for First Responders
• Deep Consensus Learning
• Constrained Convolutional Neural Networks for Weakly Supervised Segmentation
• A Survey of FPGA-Based Robotic Computing
• The Ethical Dilemma when (not) Setting up Cost-based Decision Rules in Semantic Segmentation
• Automatic Vertebra Localization and Identification in CT by Spine Rectification and Anatomically-constrained Optimization
• The Effects of Image Pre- and Post-Processing, Wavelet Decomposition, and Local Binary Patterns on U-Nets for Skin Lesion Segmentation
• Taming Adversarial Domain Transfer with Structural Constraints for Image Enhancement
• Understand Scene Categories by Objects: A Semantic Regularized Scene Classifier Using Convolutional Neural Networks
• Multi-scale GCN-assisted two-stage network for joint segmentation of retinal layers and disc in peripapillary OCT images
• Artificial Intelligence in Tumor Subregion Analysis Based on Medical Imaging: A Review
• Target-Aware Object Discovery and Association for Unsupervised Video Multi-Object Segmentation
• Feature Selective Networks for Object Detection
• Hi-Fi: Hierarchical Feature Integration for Skeleton Detection
• Robust Optimization for Deep Regression
• What Can Help Pedestrian Detection?
• Convolutional Channel Features
• Recurrent Multimodal Interaction for Referring Image Segmentation
• Beyond Planar Symmetry: Modeling human perception of reflection and rotation symmetries in the wild
• BPGrad: Towards Global Optimality in Deep Learning via Branch and Pruning
• Image Completion on CIFAR-10
• Scene Parsing via Dense Recurrent Neural Networks with Attentional Selection
• StuffNet: Using 'Stuff' to Improve Object Detection
• Tube-CNN: Modeling temporal evolution of appearance for object detection in video
• Exploiting Inter-pixel Correlations in Unsupervised Domain Adaptation for Semantic Segmentation
• Generator evaluator-selector net for panoptic image segmentation and splitting unfamiliar objects into parts
• Deep, Dense, and Low-Rank Gaussian Conditional Random Fields
• Automatic Latent Fingerprint Segmentation
• C-DLinkNet: considering multi-level semantic features for human parsing
• BABO: Background Activation Black-Out for Efficient Object Detection
• Perceive Where to Focus: Learning Visibility-aware Part-level Features for Partial Person Re-identification
• Modelling the Scene Dependent Imaging in Cameras with a Deep Neural Network
• TLGAN: document Text Localization using Generative Adversarial Nets
• Augmentation Inside the Network
• EDDA: Explanation-driven Data Augmentation to Improve Explanation Faithfulness
• Towards Understanding the Effectiveness of Attention Mechanism
• Frequency Pooling: Shift-Equivalent and Anti-Aliasing Downsampling
• Now that I can see, I can improve: Enabling data-driven finetuning of CNNs on the edge
• Depth Assisted Full Resolution Network for Single Image-based View Synthesis
• Using Cross-Model EgoSupervision to Learn Cooperative Basketball Intention
• Gaussian Filter in CRF Based Semantic Segmentation
• Triplet-based Deep Similarity Learning for Person Re-Identification
• Image2song: Song Retrieval via Bridging Image Content and Lyric Words
• Learning to Segment Object Candidates via Recursive Neural Networks
• Encoder-Decoder based CNN and Fully Connected CRFs for Remote Sensed Image Segmentation
• Learning Neural Parsers with Deterministic Differentiable Imitation Learning
• Automating Carotid Intima-Media Thickness Video Interpretation with Convolutional Neural Networks
• Vehicle Image Generation Going Well with The Surroundings
• Using Satellite Imagery for Good: Detecting Communities in Desert and Mapping Vaccination Activities
• Track Facial Points in Unconstrained Videos
• Architecture-aware Network Pruning for Vision Quality Applications
• Character Proposal Network for Robust Text Extraction
• CloudifierNet -- Deep Vision Models for Artificial Image Processing
• Temporally Folded Convolutional Neural Networks for Sequence Forecasting
• A Study on Trees's Knots Prediction from their Bark Outer-Shape
• A Survey On 3D Inner Structure Prediction from its Outer Shape
• Authoring image decompositions with generative models
• RethNet: Object-by-Object Learning for Detecting Facial Skin Problems
• Rethinking Fully Convolutional Networks for the Analysis of Photoluminescence Wafer Images
• Visual Descriptor Learning from Monocular Video
• Split-Merge Pooling
• DETCID: Detection of Elongated Touching Cells with Inhomogeneous Illumination using a Deep Adversarial Network
• Learning Pixel Representations for Generic Segmentation
• Block-Cyclic Stochastic Coordinate Descent for Deep Neural Networks