Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

min-st-cut interface #56

Open
DerThorsten opened this issue May 21, 2017 · 1 comment
Open

min-st-cut interface #56

DerThorsten opened this issue May 21, 2017 · 1 comment
Milestone

Comments

@DerThorsten
Copy link
Owner

implement min-st-cut interface similar to multicut/lifted multucut

@DerThorsten DerThorsten added this to the v1.0 milestone May 21, 2017
@DerThorsten
Copy link
Owner Author

DerThorsten commented Jun 22, 2017

@ahmadnish you can look into this.
The goal is to implement minstcut (also known as GraphCut ) functionality similar to multicut, mincut, lifted_multicut.
In fact, mincut and minstcut are very simmilar.
you just need to add unaries to mincut.
This just implements a Potts Graphical model with binary variables.
Do you think you can handle this?

One Needs to implent:

  • MinStCutObjective
  • MinStCutFactoryBase
  • MinStCutFactory
  • MinStCutBase
  • Visitors
  • If you are very very brave you can also try to implement a wrapper to
    a kolmogorov maxflow implementation but consider this task to be very optional

For most of these things one should derive from the matching base classes.
Basically look into the mincut and do the sane thing for a minstcut.

For mincut, mulicut, and lifted_multicut one just needs weights which are the beta of the potts regularizer.
For a minStCut we need a NodeMap which stores a std::pair<float,float> (or simmilar).
This can be realized by adding an other template to MinStCutObjective in addition to the WEIGHT_TYPE (the edge weights).
The name of the template could be UNARIES

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant