Data

Tools

Indexes

Applications

Experiments

Open Source Science

u-segformer-hyper

Seismic Facies Segmentation via A Segformer-based Specific Encoder-Decoder-Hypercolumns Scheme

https://github.com/emailwzg/u-segformer-hyper

Science Score: 67.0%

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
    Found 2 DOI reference(s) in README
  • ✓
    Academic publication links
    Links to: zenodo.org
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (5.9%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·

Repository

Seismic Facies Segmentation via A Segformer-based Specific Encoder-Decoder-Hypercolumns Scheme

Basic Info
  • Host: GitHub
  • Owner: emailwzg
  • License: mit
  • Language: Jupyter Notebook
  • Default Branch: main
  • Homepage:
  • Size: 6.28 MB
Statistics
  • Stars: 2
  • Watchers: 1
  • Forks: 0
  • Open Issues: 1
  • Releases: 1
Created over 3 years ago · Last pushed about 3 years ago
Metadata Files
Readme License Citation

README.md

DOI

Seismic Facies Segmentation via A Segformer-based Specific Encoder-Decoder-Hypercolumns Scheme

The paper is available in [https://doi.org/10.1109/TGRS.2023.3244037]

Installation

Training and testing can run four files directly (patchtrain.py，patchtest.py，sectiontrain.py，sectiontest.py). code file: core/models/USegformerHyper.py：Network structure of U-Segformer-Hyper、U-Segformer、Segformer core/models/patchdeconvnet.py：Network structure of patch-based Benchmark model core/models/sectiondeconvnet.py：Network structure of section-based Benchmark model

patchtrain.py：U-Segformer-Hyper,U-Segformer and Segformer are trained by patch-based mode patchtest.py：U-Segformer-Hyper,U-Segformer and Segformer are tested by patch-based mode

sectiontrain.py：U-Segformer-Hyper,U-Segformer and Segformer are trained by section-based mode sectiontest.py：U-Segformer-Hyper,U-Segformer and Segformer are tested by section-based mode

inline200-xline300.ipynb：Used to draw prediction sections metricget-5-5.ipynb：Used to draw confusion matrix

data file: runs-section：Saves the predicted data of section-based mode runs-patch：Saves the predicted data of patch-based mode

data：Training set and test set

inline：Predicted inline sections xline：Predicted xline sections

confusion_55：Confusion matrix for the test set

=======================Chinese version 中文说明=========

训练和测试直接运行patchtrain.py，patchtest.py，sectiontrain.py，sectiontest.py四个文件即可。 代码说明： core/models/USegformerHyper.py：U-Segformer-Hyper、U-Segformer、Segformer的网络结构 core/models/patchdeconvnet.py：基于patch模式的Benchmark网络 core/models/sectiondeconvnet.py：基于section模式的Benchmark网络

patchtrain.py：基于patch模式训练的U-Segformer-Hyper、U-Segformer、Segformer网络 patchtest.py：用基于patch模式训练的U-Segformer-Hyper、U-Segformer、Segformer网络进行测试

sectiontrain.py：基于section模式训练的U-Segformer-Hyper、U-Segformer、Segformer网络 sectiontest.py：用基于section模式训练的U-Segformer-Hyper、U-Segformer、Segformer网络进行测试

inline200-xline300.ipynb：用于绘制预测剖面 metricget-5-5.ipynb：绘制测试集的混淆矩阵

数据： runs-section：保存基于section模式的预测数据 runs-patch：保存基于patch模式的预测数据

data：训练集和测试集

inline：预测的inline剖面 xline：预测的xline剖面

confusion_55：测试集的混淆矩阵

Owner

  • Name: Zhiguo Wang
  • Login: emailwzg
  • Kind: user
  • Location: Xi'an, China
  • Company: Xi'an Jiaotong University

Professor

Citation (CITATION.cff) 

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
authors:
- family-names: "Wang"
  given-names: "Zhiguo"
  orcid: "https://orcid.org/0000-0003-0343-7278"
- family-names: "Wang"
  given-names: "Qiannan"
  orcid: ""
- family-names: "Chen"
  given-names: "Yumin"
  orcid: ""
title: "U-Segformer-Hyper"
version: 0.1
doi: 10.5281/zenodo.7199187
date-released: 2022-10-14
url: "https://github.com/emailwzg/U-Segformer-Hyper"

GitHub Events

Total
  • Watch event: 1
Last Year
  • Watch event: 1