Skip to content

Source code for the paper A New Approach for Evaluating and Improving the Performance of Segmentation Algorithms on Hard-to-Detect Blood Vessels by João Pedro Parella, Matheus Viana da Silva and Cesar Henrique Comin.

Notifications You must be signed in to change notification settings

jpparella/vessel_salience

Repository files navigation

Blood vessel salience quantification and augmentation

Source code for the paper A New Approach for Evaluating and Improving the Performance of Segmentation Algorithms on Hard-to-Detect Blood Vessels by João Pedro Parella, Matheus Viana da Silva and Cesar Henrique Comin.

The script vessel_salience/salience.py can be used for calculating the local vessel salience (LVS) index and the low-salience recall (LSRecall) using the API functions lvs and ls_recall.

The script vessel_salience/augmentation.py can be used for augmenting blood vessel segments using the function create_image.

Required packages are indicated in the file requirements.txt.

The Jupyter notebooks in the folder notebooks show examples for running the scripts. They are still a work in progress but show basic functionality.

About

Source code for the paper A New Approach for Evaluating and Improving the Performance of Segmentation Algorithms on Hard-to-Detect Blood Vessels by João Pedro Parella, Matheus Viana da Silva and Cesar Henrique Comin.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •