Ilastik workflow autocontext

Hey folks,

does anyone here have experience with the use of the wokflow ‘Autocontext’ from Ilastik?
I segmented a small series of images there (after visual inspection, the segmentation improved compared to simple pixel classification workflow). Nevertheless, I wanted to compare the Out of bag Errror (Obb) for comparison, but when I want to check a specific image, I get the following error message:

Does anyone have a solution?

Hello @hofmanpa,

sorry to see that autocontext is misbehaving for you. Can you readily reproduce this error? It would be very helpful to see the contents of the log. Could you maybe 1) delete the log file (`C:\Users<yourUserName>\ilastik_log.txt) and then 2) reproduce the error, and copy paste the contents of the log file here?

Cheers
Dominik

The content from the log file, the error is reproducible:

INFO 2021-03-30 18:07:39,150 ilastik_main 19124 20588 Starting ilastik from "C:\Program Files\ilastik-1.3.3post3".
WARNING 2021-03-30 18:07:44,318 opConservationTracking 19124 20588 Could not find any ILP solver
WARNING 2021-03-30 18:07:44,365 opStructuredTracking 19124 20588 Could not find any ILP solver
WARNING 2021-03-30 18:07:44,369 structuredTrackingWorkflow 19124 20588 Could not find any learning solver. Tracking will use flow-based solver (DPCT). Learning for tracking will be disabled!
INFO 2021-03-30 18:07:49,719 projectManager 19124 20588 Opening Project: C:/Users/paulh/Dokumente/Promotion/Dataset_Image_preprocessing/Autocontext_Finevo_Segmentation.ilp
WARNING 2021-03-30 18:07:49,730 warnings 19124 20588 dataset.py(313): H5pyDeprecationWarning: dataset.value has been deprecated. Use dataset[()] instead.
INFO 2021-03-30 18:07:50,534 acceleratesupport 19124 20588 No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'
WARNING 2021-03-30 18:07:51,153 warnings 19124 20588 slot.py(1287): UserWarning: Deprecated use slot.operator property instead
WARNING 2021-03-30 18:07:52,315 warnings 19124 20588 opSimpleBlockedArrayCache.py(43): FutureWarning: Conversion of the second argument of issubdtype from `dtype` to `np.generic` is deprecated. In future, it will be treated as `np.object_ == np.dtype(dtype).type`.
WARNING 2021-03-30 18:07:52,324 warnings 19124 20588 opSlicedBlockedArrayCache.py(141): FutureWarning: Conversion of the second argument of issubdtype from `dtype` to `np.generic` is deprecated. In future, it will be treated as `np.object_ == np.dtype(dtype).type`.
WARNING 2021-03-30 18:07:54,792 warnings 19124 20588 tiling.py(366): UserWarning: FIXME: This is a slow way to look for the items we want.
_TilesCache._layerCache should be a dict-of-dict-of-dict for faster lookup!
WARNING 2021-03-30 18:08:58,850 warnings 19124 20588 _methods.py(135): RuntimeWarning: Degrees of freedom <= 0 for slice
WARNING 2021-03-30 18:08:58,851 warnings 19124 20588 _methods.py(105): RuntimeWarning: invalid value encountered in true_divide
WARNING 2021-03-30 18:08:58,851 warnings 19124 20588 _methods.py(127): RuntimeWarning: invalid value encountered in true_divide
ERROR 2021-03-30 18:08:58,851 excepthooks 19124 20588 Unhandled exception in thread: 'MainThread'
ERROR 2021-03-30 18:08:58,854 excepthooks 19124 20588 Traceback (most recent call last):
  File "C:\Program Files\ilastik-1.3.3post3\ilastik-meta\ilastik\ilastik\applets\pixelClassification\FeatureSelectionDialog.py", line 810, in _run_selection
    self.opFilterFeatureSelection.setupOutputs()
  File "C:\Program Files\ilastik-1.3.3post3\ilastik-meta\ilastik\ilastik\applets\pixelClassification\opPixelClassification.py", line 707, in setupOutputs
    data, labels.astype("int"), self._filter_method
  File "C:\Program Files\ilastik-1.3.3post3\lib\site-packages\ilastik_feature_selection\filter_feature_selection.py", line 47, in __init__
    self._X = normalize_data_for_MI(np.asarray(X))
  File "C:\Program Files\ilastik-1.3.3post3\lib\site-packages\ilastik_feature_selection\filter_feature_selection.py", line 44, in normalize_data_for_MI
    X[:, i] -= X[:, i].min()
  File "C:\Program Files\ilastik-1.3.3post3\lib\site-packages\numpy\core\_methods.py", line 29, in _amin
    return umr_minimum(a, axis, None, out, keepdims)
ValueError: zero-size array to reduction operation minimum which has no identity

So far it has always been the third picture of four

thank you for providing the log!

aaah. Are you clicking on suggest features? Looks like you need to have annotations on your image in order to use it (sorry, this is an issue with our UI in the lesser used parts of ilastik). I’ll look into it some more!

Hi @k-dominik ,

I used the autocontext as in the documentation or in the published article by Kreshuk A. et al and then determined the obb error via ‘Suggested features’ after the second training applet. In some images it worked smoothly and in others I got the error message described above.
Thanks!