Fully Connected Deep Structured Networks

Abstract

Convolutional neural networks with many layers have recently been shown to achieve excellent results on many high-level tasks such as image classification, object detection and more recently also semantic segmentation. Particularly for semantic segmentation, a two-stage procedure is often employed. Hereby, convolutional networks are trained to provide good local pixel-wise features for the second step being traditionally a more global graphical model. In this work we unify this two-stage process into a single joint training algorithm. We demonstrate our method on the semantic image segmentation task and show encouraging results on the challenging PASCAL VOC 2012 dataset.

References in corpus (9)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Natural Language Processing (almost) from Scratch
• Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
• Fully Convolutional Networks for Semantic Segmentation
• Deep Captioning with Multimodal Recurrent Neural Networks (m-RNN)
• Show and Tell: A Neural Image Caption Generator
• Deep Visual-Semantic Alignments for Generating Image Descriptions
• From Captions to Visual Concepts and Back
• Fast Image Scanning with Deep Max-Pooling Convolutional Neural Networks

Cited by in corpus (74)

• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• Conditional Random Fields as Recurrent Neural Networks
• ParseNet: Looking Wider to See Better
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• From Big to Small: Multi-Scale Local Planar Guidance for Monocular Depth Estimation
• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Semantic Image Segmentation via Deep Parsing Network
• Non-local Neural Networks
• Adversarial Networks for the Detection of Aggressive Prostate Cancer
• Learning Deep Structured Models
• Gated-SCNN: Gated Shape CNNs for Semantic Segmentation
• Efficient piecewise training of deep structured models for semantic segmentation
• Semantic Facial Expression Editing using Autoencoded Flow
• Detecting Visual Relationships with Deep Relational Networks
• Learning to Refine Object Segments
• Semi and Weakly Supervised Semantic Segmentation Using Generative Adversarial Network
• Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
• Training Deep Networks with Structured Layers by Matrix Backpropagation
• Learning deep structured active contours end-to-end
• Combining the Best of Convolutional Layers and Recurrent Layers: A Hybrid Network for Semantic Segmentation
• Deep Learning for Semantic Part Segmentation with High-Level Guidance
• Semantic Image Segmentation with Task-Specific Edge Detection Using CNNs and a Discriminatively Trained Domain Transform
• Monocular Object Instance Segmentation and Depth Ordering with CNNs
• Context-self contrastive pretraining for crop type semantic segmentation
• Creativity: Generating Diverse Questions using Variational Autoencoders
• Stacked U-Nets: A No-Frills Approach to Natural Image Segmentation
• Monocular Depth Estimation using Multi-Scale Continuous CRFs as Sequential Deep Networks
• Deeply Learning the Messages in Message Passing Inference
• Video Propagation Networks
• Learning Affinity via Spatial Propagation Networks
• ModaNet: A Large-Scale Street Fashion Dataset with Polygon Annotations
• InstanceCut: from Edges to Instances with MultiCut
• Deep Learning Markov Random Field for Semantic Segmentation
• Instance-Level Segmentation for Autonomous Driving with Deep Densely Connected MRFs
• Machine learning for metal additive manufacturing: Predicting temperature and melt pool fluid dynamics using physics-informed neural networks
• Graph Convolutional Networks for Classification with a Structured Label Space
• Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes
• Structured Prediction Energy Networks
• Frank-Wolfe Network: An Interpretable Deep Structure for Non-Sparse Coding
• Exploring Context with Deep Structured models for Semantic Segmentation
• MaskRNN: Instance Level Video Object Segmentation
• DSDNet: Deep Structured self-Driving Network
• Learning Deep Representations for Semantic Image Parsing: a Comprehensive Overview
• Learning Structured Inference Neural Networks with Label Relations
• Combinatorial Energy Learning for Image Segmentation
• Learning Person Trajectory Representations for Team Activity Analysis
• Asynchronous Temporal Fields for Action Recognition
• Nonlinear Markov Random Fields Learned via Backpropagation
• Predict and Constrain: Modeling Cardinality in Deep Structured Prediction
• An End-to-end Approach to Semantic Segmentation with 3D CNN and Posterior-CRF in Medical Images
• Deep Gaussian Conditional Random Field Network: A Model-based Deep Network for Discriminative Denoising
• Semantic Object Parsing with Graph LSTM
• End-to-End Training of Hybrid CNN-CRF Models for Stereo
• Value of Temporal Dynamics Information in Driving Scene Segmentation
• Structured Prediction using cGANs with Fusion Discriminator
• A Network Structure to Explicitly Reduce Confusion Errors in Semantic Segmentation
• A Projected Gradient Descent Method for CRF Inference allowing End-To-End Training of Arbitrary Pairwise Potentials
• Superpixel Convolutional Networks using Bilateral Inceptions
• Integrated Inference and Learning of Neural Factors in Structural Support Vector Machines
• Predictive Learning on Hidden Tree-Structured Ising Models
• Detect, Replace, Refine: Deep Structured Prediction For Pixel Wise Labeling
• Progressively Diffused Networks for Semantic Image Segmentation
• Deep Learning on Attributed Graphs: A Journey from Graphs to Their Embeddings and Back
• Scaling Matters in Deep Structured-Prediction Models
• Shape-aware Generative Adversarial Networks for Attribute Transfer
• Lift-the-flap: what, where and when for context reasoning
• Deep Structured Reactive Planning
• Efficient Structured Surrogate Loss and Regularization in Structured Prediction
• Efficient Continuous Relaxations for Dense CRF
• CNN in MRF: Video Object Segmentation via Inference in A CNN-Based Higher-Order Spatio-Temporal MRF
• Beyond Gradient Descent for Regularized Segmentation Losses
• Learning to segment with image-level supervision
• Variational reaction-diffusion systems for semantic segmentation