Segmentation Framework

Descrption:

• This repository describes a PDE based segmentation framework by Level set evolution.
• Currently Chanvesse , Yezzi, and Bhattacharya energy functionals are implemented, we also provide `interactive feedback-control' implementation.
• Freeform initial region selection is supported.

Project structure:

• test_driver.py - contains the basic executable lines.
• seg_fwk.py - contains the rudimentary framework with extendable driver & stub sections.
• initialize_phi.py - This module helps in distance func initialization with initial region selection.
• schemes.py - This module contain all implemented energy functionals, that can be plug and played.
• redistance.py - This module contains re-distancing PDE for Level sets. current implementation only has Sussman.

Test Runs:

• Clone the repository

•   from seg_fwk import segmentation
    
    s = segmentation(imname='twoObj',algo='yezzi', dt=0.5)
    s.execute()
  
    s = segmentation(imname='zebra',algo='bhattacharya', dt=0.5)
    s.execute()
    
    s = segmentation(imname='airplane',algo='chanvesse', dt=0.5)
    s.execute()
    
    s = segmentation(imname='airplane',algo='ctrl1', dt=0.5)
    s.execute()
    
  

• Availabe options for imname are: twoObj,zebra,airplane; Availabe options for algos are: chanvesse,yezzi,bhattacharya,ctrl1.