Edit.md

 * [Edit segmentation on single time frame](#edit-segmentation-on-single-time-frame)
 * [Edit segmentation on several time frames](#edit-segmentation-on-several-time-frames)
-* [Cleaning options](#cleaning-options)
-* [Group cells options](#group-cells-options)
-
+* [Cleaning options](#cleaning-options): general cleaning/checking of the segmentation and EpiCure files to ensure.
+* [Group cells options](#group-cells-options): classify cells into group (type) of cells.
+* [ROI options](#roi-options): apply some process to only some drawn region.
+* [Seeds options](#seeds-options): to perform a watershed segmentation from several placed seed points.:heavy_division_sign: 
 
 # Edit segmentation on single time frame
 
-`EpiCure` proposes features to ease the manual segmentation correction on single time frame (the current one), through several [shortcuts](#shortcuts) and special [features](#features).
+`EpiCure` proposes features to ease the manual segmentation correction on single time frame (the current one), through several [shortcuts](#shortcuts).
 
 
 ## Shortcuts
@@ -62,12 +63,7 @@ To replace a whole label by a new value (label), you can use the `fill` option.
 Complete list of Napari label layer shortcuts (From File>Preferences>Shortcuts):
 ![Napari_labelshortcuts](uploads/f39d473c81d1d5e9a463ab2992d65e37/Napari_labelshortcuts.png)
 
-## Features
 
-Several features are proposed:
-* [Check Map](https://gitlab.pasteur.fr/gletort/epicure/-/wikis/Check-Map) to note the cells that are correctly segmented and already verified :white_check_mark:.
-* [Suspect outliers](https://gitlab.pasteur.fr/gletort/epicure/-/wikis/Outliers) determine cells suspicious of having an error, based on a few morphological features :mag: 
-* [Seeds](https://gitlab.pasteur.fr/gletort/epicure/-/wikis/Seeds) to perform a watershed segmentation from several placed seed points.:heavy_division_sign: 
 
 <br>
 
@@ -86,6 +82,7 @@ Then, for each <kbd>left-click</kbd>, the cell below the cursor will be set to t
 
 To finish the track editing, <kbd>right-click</kbd> and a confirmation message will appear. If you reach the end of the movie and do one more <kbd>left-clik</kbd>, the manual tracking mode will also be stopped.  
 
+---
 ---
 
 # Cleaning options
@@ -129,3 +126,29 @@ A cell can be assigned to the current group by holding <kbd>Shift+Left click</kb
 ![group-ui](uploads/bc7efcc62c0b8b64b96500f4a2f3f17a/group-ui.png)
 
 When the segmentation is saved by clicking on <kbd>s</kbd> when the `Segmentation` layer is active, the group information is also saved, in the EpiCure data file (file in the `epics` folder that ends with `_epidata.pkl`. When the movie is reopened in EpiCure, it will be automatically reloaded. 
+
+# ROI options
+
+# Seeds options
+
+**This option performs an automatic segmentation of cells starting from manually placed points (seeds).**
+
+## Manual positioning of seeds
+
+To place/add seeds, you can either check the `Seed options` parameter in the `Edit` panel, press `create seeds layer` if it's not there yet and select the `+` button in the top left panel to add points. 
+
+_Shortcut_: You can directly press <kbd>e</kbd> when the `Segmentation` layer is active, and each <kbd>left-click</kbd> will place a new seed if there is no cell below. <kbd> right-click</kbd> to stop adding seeds.
+
+## Automatic positioning of seeds from previous time point
+
+In the case of movies, we can use the information from the previous frame to place the seeds in the region that are not segmented yet. This option `Load seeds from previous time point` simply takes the centers of each cell of the previous frame as a seed.
+
+## Automatic segmentation from seeds
+Several methods to segment the cells from the seeds are proposed:
+* Intensity-based: classic watershed method, starting filling from the seeds and stopping at high intensity gradient 
+* Distance-based: purely based on the distance to the seed points, voronoi-like segmentation.
+* Diffusion-based: random walk approach from the seeds. The segmentation is spread by diffusion from the seeds through similar intensity pixels in the image (sensitive to both the distance and the intensity).
+
+The parameter `max cell radius` controls the maximum cell size. It will constrained the segmentation for the `Distance-based` method and is used to crop the image around the seeds for computation efficiency with the two other methods.
+
+Select the segmentation method and click on `Segment cells from seeds` to run it.