Repository of code for my Major Qualifying Project at Worcester Polytechnic Institute. It also serves as a supplement to the paper "Permutation-Invariant Consensus over Crowdsourced Labels" by Michael Giancola, Randy Paffenroth, and Jacob Whitehill.
The dependency list for our model is in requirements.txt. If you are on a system with pip, the easiest way to install the dependencies is by running "sudo pip install -r requirements.txt"
To run the standard simulation, execute "python simulation.py ..." within the Simulation/ directory. To run the image segmentation experiments, execute "python runsegmentation.py ..." within the Simulation/ directory.
Contains all files related to running our model, including scripts, the model implementation, and test data and results.
Contains code for generating simulation data
Contains scripts and data from the ADE20K project that were used in simulations.
Graveyard of old code/results