Bayesian SegNet: Model Uncertainty in Deep Convolutional Encoder-Decoder Architectures for Scene Understanding (2015-11-09T00:00:00.000000Z)

TL;DR

A practical system which is able to predict pixel-wise class labels with a measure of model uncertainty, and shows that modelling uncertainty improves segmentation performance by 2-3% across a number of state of the art architectures such as SegNet, FCN and Dilation Network, with no additional parametrisation.

Abstract

We present a deep learning framework for probabilistic pixel-wise semantic segmentation, which we term Bayesian SegNet. Semantic segmentation is an important tool for visual scene understanding and a meaningful measure of uncertainty is essential for decision making. Our contribution is a practical system which is able to predict pixel-wise class labels with a measure of model uncertainty. We achieve this by Monte Carlo sampling with dropout at test time to generate a posterior distribution of pixel class labels. In addition, we show that modelling uncertainty improves segmentation performance by 2-3% across a number of state of the art architectures such as SegNet, FCN and Dilation Network, with no additional parametrisation. We also observe a significant improvement in performance for smaller datasets where modelling uncertainty is more effective. We benchmark Bayesian SegNet on the indoor SUN Scene Understanding and outdoor CamVid driving scenes datasets.

Authors

Alex Kendall

8 papers

Vijay Badrinarayanan

7 papers

R. Cipolla

17 papers

TL;DR

Abstract

Authors

References48 items

Multi-task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics

What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?

The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation

Multi-Scale Context Aggregation by Dilated Convolutions

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation

Modelling uncertainty in deep learning for camera relocalization

SUN RGB-D: A RGB-D scene understanding benchmark suite

Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Robust Semantic Pixel-Wise Labelling

Learning Deconvolution Network for Semantic Segmentation

Conditional Random Fields as Recurrent Neural Networks

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs

Hypercolumns for object segmentation and fine-grained localization

Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture

Fully convolutional networks for semantic segmentation

Very Deep Convolutional Networks for Large-Scale Image Recognition

Learning Rich Features from RGB-D Images for Object Detection and Segmentation

Neural Decision Forests for Semantic Image Labelling

Caffe: Convolutional Architecture for Fast Feature Embedding

Microsoft COCO: Common Objects in Context

Visualizing and Understanding Convolutional Networks

Indoor Semantic Segmentation using depth information

ImageNet classification with deep convolutional neural networks

Indoor Segmentation and Support Inference from RGBD Images

Local Label Descriptor for Example Based Semantic Image Labeling

RGB-(D) scene labeling: Features and algorithms

Practical Variational Inference for Neural Networks

Structured class-labels in random forests for semantic image labelling

Real-time human pose recognition in parts from single depth images

Superparsing

Semantic Segmentation of Urban Scenes Using Dense Depth Maps

What, Where and How Many? Combining Object Detectors and CRFs

Label propagation in video sequences

Scene Text Recognition using Higher Order Language Priors

Semantic object classes in video: A high-definition ground truth database

Segmentation and Recognition Using Structure from Motion Point Clouds

SIFT Flow: Dense Correspondence across Different Scenes

Semantic texton forests for image categorization and segmentation

Active Learning with Statistical Models

A Practical Bayesian Framework for Backpropagation Networks

Transforming Neural-Net Output Levels to Probability Distributions

International Journal of Computer Vision

Dropout: a simple way to prevent neural networks from overfitting

Combining Appearance and Structure from Motion Features for Road Scene Understanding

TextonBoost for Image Understanding: Multi-Class Object Recognition and Segmentation by Jointly Modeling Texture, Layout, and Context

Edinburgh Research Explorer The PASCAL Visual Object Classes (VOC) Challenge

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names