Puzzle_Imaging

Code for Puzzle Imaging Manuscript

A very brief description of the files is below. Further comments are found within the files. If you have questions, please email j-glaser@u.northwestern.edu

Voxel_Puzzling

• vox_puzzling_main.m

  • This script does all of the voxel puzzling demonstrations from the manuscript (plotting and accuracy w/ different parameters)
  • Uses vox_puzzling_func.m

• vox_puzzling_func.m

  • This is the main function for voxel puzzling, and does all the steps
  • Uses simulate_neur_and_vox.m, uli_func.m, sdm_func.m, rotate_match_func.m, scomponents.m

• simulate_neur_and_vox.m

  • This function runs the basic simulation to determine where the voxels are, and which neurons are in which voxels.
  • Uses neur_per_vox_func.m

• neur_per_vox_func.m

  • For a vector of voxel sizes, says how many neurons there should be to completely fill the voxels.
  • Based off results of neurpervox.m

• neurpervox.m

  • Script that says how many neurons are necessary to fill a voxel

• plot_fig3.m

  • Script to make results plots in Figure 3

• Data_files

  • Contains the results used for plotting

Connectomics_Puzzling

• conn_puzzling_main.m

  • This script will do all of the connectomics puzzling applications from the manuscript.
  • Uses conn_puzzling_func.m

• conn_puzzling_func.m

  • This is the main connectomics puzzling function
  • Uses simulate_neur_and_conn, simulate_neur_and_conn_layers, sdm_func.m, rotate_match_func.m, scomponents.m

• simulate_neur_and_conn.m

  • This function simulates neuron locations and the connections between neurons given a single connection probability function (a single layer)

• simulate_neur_and_conn_layers.m

  • This function simulates neuron locations and the connections between neurons given multiple connection probability functions (two layers)

• plot_fig5.m

  • Script to make results plots in Figure 3

• Data_files

  • Contains the results used for plotting

Chemical_Puzzling

• chem_puzzling_main.m

  • This script produces all the figure outputs for the chemical puzzling applications
  • Uses chem_puzzling_func.m
  • Uses the simulations produced from chem_sim.m

• chem_puzzling_func.m

  • This is the main function for chemical puzzling.
  • Uses uli_func.m, rotate_func_2d, scomponents.m

• chem_sim.m

  • This function does the chemical puzzling simulation that the results in the paper come from.
  • Uses P3.jpg, the image representing the chemical concentration of the plate

• rotate_func_2d.m

  • This function rotates the reconstructed position and gives the sum squared error between this rotation and the true position. It works for 2 dimensional inputs.

• Data_files

  • Contains the simulation outputs

General_Files

• uli_func.m

  • This function does the ULI (unweighted landmark isomap) algorithm to reduce dimensions
  • Uses dist_from_pt2.m

• dist_from_pt2.m

  • This function outputs the geodesic distance of all points to a given landmark point

• sdm_func.m

  • This function does the SDM (sparse diffusion maps) algorithm to reduce dimensions

• rotate_match_func.m

  • This function finds the rotation (and potential reflection) to best match the reconstructed position to the true position.
  • Uses rotate_func.m

• rotate_func.m

  • This function rotates the reconstructed position and gives the sum squared error between this rotation and the true position.

• scomponents.m

  • Finds the connected components of a graph
  • Uses sparse_to_csr.m