M-DenseUNet: Multi Dense Encoder Connected UNet for Biomedical Image Segmentation

In biomedical image segmentation, the desired performance is necessary for more detailed segmentation. In this paper, we propose the M-DenseUNet which combines multi Dense Encoders and U-Net. The encoder of M-DenseUNet consists of convolutional layers, Dense Block and Transition. We show that such a...

Full description

Saved in:
Bibliographic Details
Published in2022 IEEE 11th Global Conference on Consumer Electronics (GCCE) pp. 919 - 921
Main Authors Jin, Tongdan, Chen, Kaixu, Yamane, Satoshi, Kuroda, Yoshihiro
Format Conference Proceeding
LanguageEnglish
Published IEEE 18.10.2022
Subjects
Biomedical image segmentation
Biomedical imaging
Consumer electronics
Dense Encoder
DenseNet
Feature extraction
Image segmentation
Task analysis
Transition
U-Net
Online AccessGet full text

Cover

More Information
Summary:In biomedical image segmentation, the desired performance is necessary for more detailed segmentation. In this paper, we propose the M-DenseUNet which combines multi Dense Encoders and U-Net. The encoder of M-DenseUNet consists of convolutional layers, Dense Block and Transition. We show that such a network can strengthen feature propagation and encourage feature reuse to get details.
DOI:10.1109/GCCE56475.2022.10014305