Skip to content

mitseng/RetinalVesselSegmentation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

57 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RetinalVesselSegmentation

Retinal blood vessel segmentation based on U-Net. The model is a U-Net, We replace Residual Double Convolution Unit with Multi-Fiber Unit, which reduces the model parameters by 76.1%, while performance remains basically unchanged.

Data

Two open Dataset are used: DRIVE and CHASE_DB, and calssical preprocessing are applied: Gray-scale Conversion, Stadardization, CLAHE and γ-adjustment. We extract 64 * 64 pix size patches from DRIVE, 128 * 128 for CHASE_DB, then rotate patch to get 4 times more data, then normalize input data for model.

Model

Models can be found in src/U_Net2.py and src/MF_UNet2.py, proposed model is located in the latter file. Multi-Fiber Unit is defined in src/MF_UNet2.py. It is based on grouped convolution, multiplexers are applied to ficilitate information exchange between fibers. Multi-Fiber Unit Trained model is given in /ParameterFile.

Result

Proposed model is trained on both dataset, classical U-Net is trained on DRIVE dataset. This an image from test set of DRIVE and CHASE_DB respectively. DRIVE CHASE_DB Various metrics are used to compare the performence of two models, the most advantage of MF-U-Net is parameter size.

Model Parameter File IOU F1-Score
U-Net 32.6MB 0.6850 0.8131
MF-U-Net 7.6MB 0.6837 0.8121

About

Retinal blood vessel segmentation based on U-Net.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages