doc: save for training func no longer exists

removed documentation entry as well as related images.

README.md

@@ -91,20 +91,21 @@ The plugin uses (lossless) gzip compression.
 
 ![ImageJ Menu](./doc/screenshots/IJ-Export.png)
 
-### How to train an ilastik project to be used with those wrappers
-
-For the workflow wrappers below, it is important that ilastik projects are trained from data that was 
-preprocessed and exported the same way as all further datasets will be. There are two ways how this can be achieved with this plugin:
-
-1. You can manually use the export option as described [above](#export) to export your images from ImageJ to an ilastik-compatible 
-   HDF5 dataset with 5 dimensions. Then when you create an ilastik project, read the raw data / probabilities / segmentation from the exported HDF5 files.
-2. Each workflow wrapper has a `Save temporary file for training only, no prediction` option. If you select this option and run the plugin, 
-   the input files -- that would otherwise be passed to ilastik -- are exported to temporary files. 
-   ![Save temporary file checkbox](./doc/screenshots/IJ-Train-Checkbox.png)
-   The locations and names of those files are written to ImageJ's console, which you should open using `Window->Console`.
-   ![Temporary file location in console](./doc/screenshots/IJ-Train-Console.png)
-   Use this file as input in your ilastik project. Then processing any further files from ImageJ through the ilastik workflow wrappers
-   should give the desired results.
+### How to train an ilastik project for use in the plugins
+
+In general you can use any project of the supported workflows that you trained.
+However, there are some things to keep in mind:
+  * The spatial dimensions (not the size along those) need to match.
+    Meaning if you trained your project with 3D (`x`,`y`,`z`) data, you can only process 3D images with it.
+  * The largest 3D filter you selected in ilastik, must fit inside the image.
+    It's roughly the largest 3D sigma you selected times 3.5 (so with default largest sigma `10.0` it would be a minimum of `35` pixels in `z`-dimension).
+  * The number of channels needs to match (also they should be in the same order), if you train e.g. on dapi + fitc, you should only process files with dapi + fitc channels.
+
+
+Ideally, you train your ilastik already with `.h5` files manually generated by the [above export functionality](#export).
+The export creates files with maximum compatibility with ilastik - and you will also get the best performance during training.
+With this you make sure that the whole processing chain will work.
+
 
 ### ilastik configuration of the workflow wrappers
 Found at `Plugins -> ilastik -> Configure ilastik executable location`.