Problem with do_elastix_registration.py

[meruicecream]

Hi！I have met a problem when i want to run the do_elastix_registration.py using Pycharm. Traceback (most recent call last): File "D:/RegUn-master/do_elastix_registration.py", line 154, in do_elastix_registration() File "D:/RegUn-master/do_elastix_registration.py", line 105, in do_elastix_registration reg_elx.write_and_submit_job(setting, job_name=job_name, phase=phase, cn=cn, out=out, script_address=backup_script_address) File "D:\RegUn-master\Functions\Python\registration_elx\registration_elx.py", line 262, in write_and_submit_job queue_info, job_info = sungrid.qstat() File "D:\RegUn-master\Functions\Python\sungrid_utils_tools.py", line 20, in qstat xml = qstat2xml(qstat_path=qstat_path) File "D:\RegUn-master\Functions\Python\sungrid_utils_tools.py", line 37, in qstat2xml qstatxml = sp.check_output([qstat_path, xml_option], stderr=sp.STDOUT) File "C:\Users\13649\AppData\Local\conda\conda\envs\pytorch\lib\subprocess.py", line 356, in check_output *kwargs).stdout File "C:\Users\13649\AppData\Local\conda\conda\envs\pytorch\lib\subprocess.py", line 423, in run with Popen(popenargs, **kwargs) as process: File "C:\Users\13649\AppData\Local\conda\conda\envs\pytorch\lib\subprocess.py", line 729, in init restore_signals, start_new_session) File "C:\Users\13649\AppData\Local\conda\conda\envs\pytorch\lib\subprocess.py", line 1017, in _execute_child startupinfo) FileNotFoundError: [WinError 2] The system cannot find the file specified

What can i do to solve this problem? By the way,in chest_segmentation.py,where can i achieve ptpulmo.exe and PulmoDefaultSettings.psf? Thanks a lot

[hsokooti]

Hi, with the default settings of the code, it will run over an Oracle Grid Engine (previously Sun Grid Engine) cluster. If that is installed on your cluster, the default command "qstat" should work (qstat DOC)

To easily debug the code, maybe it could be better to run it in the "local" mode first by changing the do_elastix_registration.py#13 to:

where_to_run = 'local'

Unfortunately, the software "ptpulmo.exe" is not publicly available. You can contact Prof. Berend Stoel if you need a license for that. All in all, even a primitive segmentation would suffice for this purpose.

[meruicecream]

It is exciting to read your article. It seems that i cannot find codes about regression forest, where can i find them ? Thanks a lot!

[hsokooti]

Thank you for your interest in our article. That part is not in this repository, but the code snippet is:

from sklearn.ensemble import RandomForestRegressor

estimator = RandomForestRegressor(random_state=0,
                                  criterion='mse',
                                  n_estimators=number_of_tree,
                                  max_features=m,
                                  oob_score=False,
                                  max_depth=tree_depth,
                                  min_samples_split=min_samples_split,
                                  n_jobs=7)
estimator.fit(train_feature, train_tre)
feature_importance = estimator.feature_importances_
test_predict_tre = estimator.predict(test_feature)