segdeformer

Official implementation of SegDeformer.

https://github.com/lygsbw/segdeformer

Science Score: 44.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
    Found CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (10.9%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·

Repository

Official implementation of SegDeformer.

Basic Info
  • Host: GitHub
  • Owner: lygsbw
  • License: apache-2.0
  • Language: Python
  • Default Branch: main
  • Homepage:
  • Size: 9.22 MB
Statistics
  • Stars: 9
  • Watchers: 1
  • Forks: 1
  • Open Issues: 2
  • Releases: 0
Created over 3 years ago · Last pushed about 3 years ago
Metadata Files
Readme License Citation

README.md

A Transformer-based Decoder for Semantic Segmentation with Multi-level Context Mining

Official implementation of the paper "A Transformer-based Decoder for Semantic Segmentation with Multi-level Context Mining",

by Bowen Shi, Dongsheng Jiang, Xiaopeng Zhang, Han Li, Wenrui Dai, Junni Zou, Hongkai Xiong, Qi Tian.

[Paper] [Appendix]

framework

Installation

Our code is based on MMSegmentation. For install and data preparation, please refer to the guidelines in MMSegmentation.

Training

Example: train SegFormer-B1 + SegDeformer on ADE20K:

python startlocaltrain.py --configfile segformer/segformermit-b1512x512160kade20ksegdeformer3.py

Results

ADE20K

| Method| Backbone | Crop Size | Lr schd | mIoU | config | log | | ---------------- | -------- | --------- | -----| ----- | --------- | --------- | SegFormer-B1 | MiT-B1 | 512x512 | 160000 |40.97 | - | - | | SegFormer-B1 + SegDeformer | MiT-B1 | 512x512 | 160000 |44.12 | config | log | | SegFormer-B2 | MiT-B2 | 512x512 | 160000 |45.58 | - | - | | SegFormer-B2 + SegDeformer | MiT-B2 | 512x512 | 160000 | 47.34 | config | log | | SegFormer-B5 | MiT-B5 | 512x512 | 160000 |49.13 | - | - | | SegFormer-B5 + SegDeformer | MiT-B5 | 512x512 | 160000 | 50.34 | config | log |

Note:

  • We adapt our code to the latest version of MMSegmentation (v0.29.1), while the pretrained MiT models we used are still the old version provided by MMSegmentation (20210726 version) to keep consistent with our paper. Details can be found in this link.
  • The performance is sensitive to the seed values used, so the results might fluctuate.

Acknowledgement

This reposity is based on the MMSegmentation repository. Thanks for their contributions to the community.

Citation

If you find this repository/work helpful in your research, welcome to cite the paper. ``` @inproceedings{shi2022transformer, title={A Transformer-Based Decoder for Semantic Segmentation with Multi-level Context Mining}, author={Shi, Bowen and Jiang, Dongsheng and Zhang, Xiaopeng and Li, Han and Dai, Wenrui and Zou, Junni and Xiong, Hongkai and Tian, Qi}, booktitle={European Conference on Computer Vision}, pages={624--639}, year={2022}, organization={Springer} }

Owner

  • Name: min-sbw
  • Login: lygsbw
  • Kind: user

PhD student

Citation (CITATION.cff)

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
authors:
  - name: "MMSegmentation Contributors"
title: "OpenMMLab Semantic Segmentation Toolbox and Benchmark"
date-released: 2020-07-10
url: "https://github.com/open-mmlab/mmsegmentation"
license: Apache-2.0

GitHub Events

Total
  • Watch event: 2
Last Year
  • Watch event: 2

Dependencies

docker/Dockerfile docker
  • pytorch/pytorch ${PYTORCH}-cuda${CUDA}-cudnn${CUDNN}-devel build
docker/serve/Dockerfile docker
  • pytorch/pytorch ${PYTORCH}-cuda${CUDA}-cudnn${CUDNN}-devel build
requirements/docs.txt pypi
  • docutils ==0.16.0
  • myst-parser *
  • sphinx ==4.0.2
  • sphinx_copybutton *
  • sphinx_markdown_tables *
requirements/mminstall.txt pypi
  • mmcls >=0.20.1
  • mmcv-full >=1.4.4,<1.7.0
requirements/optional.txt pypi
  • cityscapesscripts *
requirements/readthedocs.txt pypi
  • mmcv *
  • prettytable *
  • torch *
  • torchvision *
requirements/runtime.txt pypi
  • matplotlib *
  • mmcls >=0.20.1
  • numpy *
  • packaging *
  • prettytable *
requirements/tests.txt pypi
  • codecov * test
  • flake8 * test
  • interrogate * test
  • pytest * test
  • xdoctest >=0.10.0 test
  • yapf * test