Multi-Scale Context Aggregation by Dilated Convolutions

Fisher Yu, Vladlen Koltun
11/23/2015
cs.CV

Official code: fyu/dilation (784 stars, MIT, last commit 2018-04-01).

Abstract

State-of-the-art models for semantic segmentation are based on adaptations of convolutional networks that had originally been designed for image classification. However, dense prediction and image classification are structurally different. In this work, we develop a new convolutional network module that is specifically designed for dense prediction. The presented module uses dilated convolutions to systematically aggregate multi-scale contextual information without losing resolution. The architecture is based on the fact that dilated convolutions support exponential expansion of the receptive field without loss of resolution or coverage. We show that the presented context module increases the accuracy of state-of-the-art semantic segmentation systems. In addition, we examine the adaptation of image classification networks to dense prediction and show that simplifying the adapted network can increase accuracy.

View on arXivView PDF

Code Implementations(1)

fyu/dilationOfficial100%
MIT

Dilated Convolution for Semantic Image Segmentation

784263Shell, PythonJan 14, 20168 years agoMIT

Similar Papers

Understanding Convolution for Semantic Segmentation

88% match1 repo6102017

Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers

87% match1 repo1.1k2020

Efficient Dense Modules of Asymmetric Convolution for Real-Time Semantic Segmentation

87% match1 repo332018

Vision Transformers: From Semantic Segmentation to Dense Prediction

87% match1 repo1.1k2022

ReSeg: A Recurrent Neural Network-based Model for Semantic Segmentation

87% match1 repo1262015

Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells

86% match1 repo1512018

Cite this paper

@article{yu2015multiscale,
  title  = {Multi-Scale Context Aggregation by Dilated Convolutions},
  author = {Fisher Yu and Vladlen Koltun},
  year   = {2015},
  eprint = {1511.07122},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1511.07122}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1511.07122.svg)](https://codeofpaper.com/paper/1511.07122)

Discussion