GitHub - qedq/NeuroProof: Suite of tools for analyzing and guiding image segmentation refinement.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 25 Commits
Algorithms		Algorithms
DataStructures		DataStructures
ImportsExports		ImportsExports
Priority		Priority
Utilities		Utilities
examples		examples
.gitignore		.gitignore
CMakeLists.txt		CMakeLists.txt
NeuroProof_test.cpp		NeuroProof_test.cpp
README.txt		README.txt
calculate_GPR.cpp		calculate_GPR.cpp
graph.json		graph.json
pythonNeuroProofPriorityInterface.cpp		pythonNeuroProofPriorityInterface.cpp
pythonNeuroProofPriorityInterface.h		pythonNeuroProofPriorityInterface.h
pythonRagInterface.cpp		pythonRagInterface.cpp
run-testragscript.sh		run-testragscript.sh
run-testscript.sh		run-testscript.sh
testragscript.py		testragscript.py
testscript.py		testscript.py

Repository files navigation

BUILD instructions:

1. make a link to libjsonlib.so in a default library directory
2. make a link to the 'json' directory in the default include directory
3. go in NeuroProof directory and mkdir 'build'
4. cd build; cmake ..; make
5. make link to build/libNeuroProofPriority.so to a default python directory to use the Python libraries

GPR Example:

Call sample program:

    ./bin/calculate_GPR <graph> <num_threads> <num_paths>

An example graph is provided as graph.json.  This file acts as documentation on how to format an input graph.

Calling ./bin/calcualte_GPR graph.json 4 1 will use 4 threads to compute the GPR for the graph.json.  '1' refers
to using K=1 paths for finding node affinity.  Currently K>1 is not supported in NeuroProof.  Details
can be found in an upcoming paper.