Menu Vasculature
----------------

Vasculature -> Binarization
~~~~~~~~~~~~~~~~~~~~~~~~~~~

This tab is meant to analyse the brain vasculature as in
https://doi.org/10.1016/j.cell.2020.01.028

.. figure:: images/vasculature/binarization.png
    :align: left

The first step consists in thresholding (“\ *binarize”*) your images.
In the case of the vasculature, you are likely working with 2 channels.
The first (raw) or *all vessels* channel and the second or *arteries*
channel. You will need to binarize both with some postprocessing like
smoothing and filling of the tubular structures before the program is
able to extract the graph representation of your vascular network.

The filling step is done in 2 ways. Through a “classical” image
filtering step and through a subsequent machine learning based step.

For both channels, select the steps you want to run.

.. container:: clearer

    .. image:: images/clearer.png

Clip range:
    The valid range for the binarization. Set the minimum based on the
    lowest intensities you see and the maximum based on the values of
    large vessels at the surface. You should set to the limit to a value
    that still excludes the halo.

.. HINT::
    For optimal values, open the stitched image and look at the pixel intensities inside
    the brain (ignore the pixels outside the brain).

Threshold:
    The threshold for the pseudo deconvolution. Voxels above this
    threshold will be added to the binarization result in the multi-path
    biniarization.

.. HINT::
    Select slightly below the clip range min.

After the channels are binarized, they need to be combined into a single
image. This is done with the **Combine channels** button.

.. NOTE::
    You can click on the **run** buttons independently and then
    **combine** or you can just select the values including the graph
    parameters described below and click **Run all** at the bottom of the
    controls window to run the whole vasculature pipeline at once.

Vasculature -> Display
~~~~~~~~~~~~~~~~~~~~~~
.. figure:: images/vasculature/binarization_display.png
    :align: left

You can see how the previous steps performed using the display tab.

Select the 2 images types and channels you want to display and click on
**Plot side by side** or **Plot overlay** for the corresponding display
type.

.. container:: clearer

    .. image:: images/clearer.png

Vasculature -> Graph Construction
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. figure:: images/vasculature/graph_construction.png
    :align: left

This tab will transform the binary image into a graph structure. Select
the steps you want to perform (the default is all enabled and should
work in most cases) and click **Build**.

.. NOTE::
    Once done, to save some RAM, you can click **Unload temp graphs** to
    remove them from RAM.
    The removed graphs will still be available but will need to be
    reloaded from the hard drive by the program upon request.

.. container:: clearer

    .. image:: images/clearer.png

Vasculature -> Arteries and vein tracing
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. figure:: images/vasculature/tracing_arteries.png
    :align: left

.. figure:: images/vasculature/tracing_veins.png
    :align: left

When the graph is built, the vertices are assigned an identity (vein,
capillary or artery) based on the labelling intensity in the two
channels. This is then refined at this step using the radius and
hysteresis thresholding.

Set the minimum radius of arteries and veins


.. container:: clearer

    .. image:: images/clearer.png

Vasculature -> Graph display
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. figure:: images/vasculature/graph_display.png
    :align: left

Once finished, you can visualise your graphs. Open the
**graph display -> graph visualization** tab and select from the
two options available:

.. container:: clearer

    .. image:: images/clearer.png

Plot region:
    - First click **load** to open the slicer and select an appropriate
        region to plot. If choosing to plot as meshes, the region should
        be limited (see below) because your computer cannot display too
        many vertices in this mode.
    - Then pick the graph step to plot.
        .. warning::
            The **raw** graph has many vertices and you should thus choose
            a tiny portion
    - Select your plot type
        .. warning::
            It is impossible to plot the whole brain in *mesh* mode. In this
            case, use the *line* mode.
    - Click **Plot region**

.. figure:: images/vasculature/pick_region.png

    Picking a slice to display

.. figure:: images/vasculature/vasculature_graph.png

    Example graph

Plot chunk
    - Pick the graph step to plot.
        .. warning::
            The **raw** graph has many vertices and you should thus choose
            a tiny portion.
    - Select your plot type
        .. warning::
            Use *line* if your region of choice is big.
    - Click **Plot chunk**. This will open a popup window with the
        tree of the brain structures which you can unfold to pick a
        region of your choice.

        .. figure:: images/vasculature/region_picker.png

    - Click **OK** and wait for your graph to appear.