search

lucasb-eyer / pydensecrf

Python wrapper to Philipp Krähenbühl's dense (fully connected) CRFs with gaussian edge potentials.
MIT License
1.94k stars 414 forks source link

Example Issue #15

Closed sqy12 closed 8 years ago

sqy12 commented 8 years ago

In your example, the labels after relabelling is ranged from 0 - 2 and the M is equal to 3. But the map computed is ranged from 1 - 2 which means there is a class gone? The result of the computation is quite wield in my case.

sqy12 commented 8 years ago

But it's not a large problem. How I can change the weight of each potential?

lucasb-eyer commented 8 years ago

These are again two different questions, please try to ask only one question per issue and prefer opening multiple issues for multiple questions.

I've clarified the answer to your first question in the first commit (1043a71) and to your second question in the second commit (3846e8a)