Skip to content

Latest commit

 

History

History
93 lines (72 loc) · 4.26 KB

README.md

File metadata and controls

93 lines (72 loc) · 4.26 KB
Raw

CFPNet-M: A Light-Weight Encoder-Decoder Based Network for Multimodal Biomedical Image Real-Time Segmentation

Result
Result
Result
Result
Result
This repository contains the implementation of a novel light-weight real-time network (CFPNet-Medicine: CFPNet-M) to segment different types of biomedical images. It is a medical version of CFPNet, and the dataset we used from top to bottom are **DRIVE, ISBI-2012, Infrared Breast, CVC-ClinicDB and ISIC 2018**. The details of CFPNet-M and CFPNet can be found here respectively.

CFPNet-M, CFPNet Paper, DC-UNet and CFPNet Code

🔥 NEWS 🔥 The pytorch version is available pytorch-version.

Architecture of CFPNet-M

CFP module

Result

CFPNet-M

Result

Dataset

In this project, we test five datasets:

  • Infrared Breast Dataset
  • Endoscopy (CVC-ClinicDB)
  • Electron Microscopy (ISBI-2012)
  • Drive (Digital Retinal Image)
  • Dermoscopy (ISIC-2018)

Usage

Prerequisities

The following dependencies are needed:

  • Kearas == 2.2.4
  • Opencv == 3.3.1
  • Tensorflow == 1.10.0
  • Matplotlib == 3.1.3
  • Numpy == 1.19.1

training

You can download the datasets you want to try, and just run: for UNet, DC-UNet, MultiResUNet, ICNet, CFPNet-M, ESPNet and ENet, the code is in the folder network. For Efficient-b0, MobileNet-v2 and Inception-v3, the code is in the main.py. Choose the segmentation model you want to test and run:

main.py

Segmentation Results of Five datasets

Result_table
Result_table

Speed and FLOPs

The code of calculate FLOPs are in main.py, you can run them after training.

Result_table

Citation

@article{lou2023cfpnet,
  title={Cfpnet-m: A light-weight encoder-decoder based network for multimodal biomedical image real-time segmentation},
  author={Lou, Ange and Guan, Shuyue and Loew, Murray},
  journal={Computers in Biology and Medicine},
  pages={106579},
  year={2023},
  publisher={Elsevier}
}

@inproceedings{lou2021cfpnet,
  title={Cfpnet: channel-wise feature pyramid for real-time semantic segmentation},
  author={Lou, Ange and Loew, Murray},
  booktitle={2021 IEEE International Conference on Image Processing (ICIP)},
  pages={1894--1898},
  year={2021},
  organization={IEEE}
}

@inproceedings{lou2021dc,
  title={DC-UNet: rethinking the U-Net architecture with dual channel efficient CNN for medical image segmentation},
  author={Lou, Ange and Guan, Shuyue and Loew, Murray H},
  booktitle={Medical Imaging 2021: Image Processing},
  volume={11596},
  pages={115962T},
  year={2021},
  organization={International Society for Optics and Photonics}
}