jaimeivancervantes/ilastikdiagnostictools

A simple test script to profile ilastik feature computation and prediction running times.

IlastikDiagnosticTools

A collection of diagnostics tools for Ilastik.

Features

• A parameter-optimization script src/paramOpt.py that runs cross-validation tests to find the optimal number of trees and features.
• A simple test script src/runningTimeProf.py that isolates feature computation and prediction in order to profile the time for each of these stages individually.
• A script trackingCsvDiagnostics.py to read and diagnose the conservation-tracking csv result files.

Instructions to Run the Parameter-Optimization Script

1. Download and install the latest versions of ilastik and lazyflow.
2. IMPORTANT: Change directory to the parent folder that contains the project (.ilp) file.
3. Run the parameter-optimization script with the following command:

>>>python paramOpt.py --run_command=<ilastik-run-command> --project-file=<ilp-project-file>

Instructions to Run the Profiling Script

1. Download and install the latest versions of ilastik and lazyflow.
2. Run the profiling script with the following command:

>>>python runningTimeProf.py <project-file> <data-file> <threads-enabled> <thread-number>

Instructions to Run the Csv Diagnostics Script

The file trackingCsvDiagnostics.py allows the user to plot different tracking parameters contained in the csv files. Here's a list of the arguments available.

Argument | Description --- | --- | --- | -h, --help | show this help message and exit --file FILE | Name of conservation-tracking csv file. --lineage-ids | Plot the lineage IDs. --tracks | Plot the tracks of all the objects in the file. --object-num | Plot the number of objects per frame. --errors | Plot the errors per frame. These errors include appearing IDs and repeated IDs. --collisions | Plot the number of collisions per frame. --bounding-rect BOUNDINGRECT | Plot the bounding rectangles for the choose frame.

1. Download and install the latest versions of ilastik and lazyflow.
2. Run the tracking diagnostics script with the test csv file with the following command:

>>>python trackingCsvDiagnostics.py --tracks