Hi. I changed app.py to support saving the project when importing and also for directly exporting to STL the surface generating using all the predefined presets for the masks. This allows to automatically generate the models from the images, without using the GUI. I thought it would be useful for other people too, that's why I'm creating this PR.

I also created a Dockerfile, so the environment can be easily replicated in any computer with Docker.

Any suggestion regarding refactorings or the API for the command line is welcome :smile:

btw, is there a way to use this command line options without opening a GUI? So it can be used through SSH or Docker without a display being set.

I tried to install Invesalius on Windows. After the setup had finished, I was able to start the program once. However, after closing it, I was not able to re-open it again. Whenever I tried to, a message saying "Failed to execute script app" pops up. Even reinstalling the program did not solve the issue. I even tried different versions of the program - the error remains. I couldn't find out anything else about it. I am using Windows 10.

Hi I recently downloaded and installed the latest release for ARM64, but it only shows the startup screen and then nothing happens. I even tried to install it with homebrew but i get a lot of error messages during the installation steps and Im not a programmer so i dont know how to solve the errormessages. Does anyone else got these problems with an apple silicon computer. Has anyone installed it successfully?

Regards

hi, @tfmoraes ,

can you support some x86 packages on windows, the Running InVesalius 3 in Windows website just x64...

i tried to use pip(or google...) but a lot of packages can not found....

think you.

Hi guys,

I'm using Windows 10 x64 without with an Intel graphics card. I'm running InVesalius master branch from the source code and installed the most recent libraries as shown in the wiki page. After I import a DICOM exam, create the surface using the threshold selection, the surface appears in the Volume panel but does not allow any mouse interaction or does not update with the CrossMove. Also, when I maximize it, it does not update the view.

In principle no errors in the command prompt, just a bunch of Deprecation Warnings.

What might be causing it? @tfmoraes recommended to update the Graphics card driver, but it is already updated and the problem still remains.

Thanks!

I am using InVesalius 3.1 for Windows, and some dicom files work, and some don't. When I hit 'OK', it crashes and says InVesalius not responding. Any help would be greatly appreciated. I will link an example data set that doesn't work, from OSF: http://www.mediafire.com/file/5emvhl8mjl466gm/osfstorage-archive+%2821%29.zip

I followed the instructions, tried both with the activated my_env env and without, but getting this error.

Trying pip install wx gives this:

ERROR: Could not find a version that satisfies the requirement wx (from versions: none)
ERROR: No matching distribution found for wx

Linux Mint 20.2 (Ubuntu 20.04) python --version: Python 3.8.8

System: Ubuntu LTS 16.04 InVesalius 3.1.1

Steps to reproduce:

• Open InVesalius
• Select 'Import medical images...' from left-side panel
• Select a folder with .nii file that was used in NexStim navigation ssytem --> Error message appears: "There are no DICOM files in the selected folder." --> InVesalius hangs with window "Loading DICOM files"

This message cannot be dismissed, although pressing Cancel seems to refresh the sceen. Pressing 'X' on top left of window disables the Cancel button. Attempting to close InVesalius by right clicking the task bar icon and selecting 'Close' displays a confirmation dialog, but nothing happens after that. Attempting to kill the program with '$ kill ' gets rid of the process when viewed with '$ ps', but does nothing to the program window.

The only way to close the window is to use 'ps' to find the line 'python app.py' and kill that.

First thanks for sharing! I follow steps here https://github.com/invesalius/invesalius3/wiki/Running-InVesalius-3-in-Windows

used: python3.7 Visual Studio (2017 community edition) the most recent commit

After run python app.py, everything was ok. I can open DICOM, after click "import", the 4 panels show up, but without images. I have to use mouse to click the scroll to make the panel show images.

The same commit on Ubuntu works fine and does not have this issue.

Could you please help on this ? Thanks.

Greetings

i have a slice set that is 2500x1400 at 47u per pixel

when i import it gives me a wierd squashed image in one of the axis

can the software only import square images?

Hi guys,

I'm having the following problem. I segmented the brain from the MRI of a subject and generated the polygonal mesh in a different combination of software. They all worked within the native MRI coordinate system.

However, when I open the MRI and generate the head surface in InVesalius, and then import the segmented brain surface by clicking in the button in the lower left corner of the interface, the surface is rendered in a different coordinate than it should be. For example, it is outside and far from the head (while the segmented brain has to be inside the head surface generated in InVesalius).

As far as I know, InVesalius follows the RAS+ coordinate system and all imported MRI/CT image when read as arrays goes through two coordinate transformations which are applied inside the module imagedata_utils.py in the method img2memmap in the following order:

# data is the 3D array representing the image
data = numpy.swapaxes(data, 0, 2)
data = numpy.fliplr(data)

This is a bit complicated to apply for a surface imported in the vtk polydata data. Do you have a straightforward solution? I thought about building the rotation matrices in vtk4x4Matrixand applying the transformations by vtkTransform, but in principle this would require a lot of thinking, manual working and testing.

P.S. When I open the original MRI and segmented surface in ParaView, they both are correctly aligned. Below are prictures of paraview and InVesalius.

Cheers!

On starting git-latest invesalius3 I get the message (repeated):

Traceback (most recent call last):
  File "/usr/lib/python3.10/site-packages/wx/lib/agw/aui/auibar.py", line 3510, in OnPaint
    self._art.DrawButton(dc, self, item, item_rect)
  File "/usr/lib/python3.10/site-packages/wx/lib/agw/aui/auibar.py", line 1008, in DrawButton
    bmp_rect, text_rect = self.GetToolsPosition(dc, item, rect)
  File "/usr/lib/python3.10/site-packages/wx/lib/agw/aui/auibar.py", line 1508, in GetToolsPosition
    bmp_rect = wx.Rect(bmp_x, bmp_y, bmp_width, bmp_height)
TypeError: Rect(): arguments did not match any overloaded call:
  overload 1: too many arguments
  overload 2: argument 1 has unexpected type 'float'
  overload 3: argument 1 has unexpected type 'float'
  overload 4: argument 1 has unexpected type 'float'
  overload 5: argument 1 has unexpected type 'float'
  overload 6: argument 1 has unexpected type 'float'

After loading DICOM images, I also get (repeated):

Traceback (most recent call last):                                                                                                      
  File "/usr/lib/python3.10/site-packages/invesalius/gui/widgets/canvas_renderer.py", line 189, in OnMouseMove
    x, y = self.viewer.get_vtk_mouse_position()
AttributeError: 'SingleImagePreview' object has no attribute 'get_vtk_mouse_position'

The application still functions(!). I was able to import a set of CT images and export to stl, but there are gui drawing issues. I could not change any masks. Most left side wx widgets did not appear until I manually mouse-overed the expected spots to render them. Most likely I am using too new or too old of a dependency.

My versions are:

$ sudo emerge sci-libs/gdcm dev-python/pyvista dev-python/imageio sci-libs/scikit-image sci-libs/nibabel dev-python/pyacvd dev-python/wxpython -pv -1

These are the packages that would be merged, in reverse order:

Calculating dependencies  ...... done!
[ebuild   R    ] dev-python/wxpython-4.0.7-r4:4.0::gentoo  USE="-test -webkit" PYTHON_TARGETS="python3_10 -python3_8 -python3_9" 0 KiB
[ebuild   R   *] dev-python/pyacvd-9999::jakeogh  PYTHON_TARGETS="python3_10 -python3_8 -python3_9" 0 KiB
[ebuild   R    ] sci-libs/nibabel-3.2.1::science  USE="-dicom -doc -test" PYTHON_TARGETS="python3_10 -python3_8 -python3_9" 0 KiB
[ebuild   R    ] sci-libs/scikit-image-0.19.3-r1::gentoo  USE="-test" PYTHON_TARGETS="python3_10 -python3_11" 0 KiB
[ebuild   R    ] dev-python/pyvista-0.36.1::jakeogh  PYTHON_SINGLE_TARGET="python3_10 -python3_9 (-python3_8%)" 0 KiB
[ebuild   R    ] sci-libs/gdcm-3.0.20::jakeogh  USE="python vtk -doc -test" PYTHON_SINGLE_TARGET="python3_10 -python3_9" 0 KiB
[ebuild   R    ] dev-python/imageio-2.23.0-r1::gentoo  USE="-test" PYTHON_TARGETS="python3_10 -python3_11" 0 KiB

To get the above results, I had to make a minor change:

diff --git a/invesalius/reader/dicom_reader.py b/invesalius/reader/dicom_reader.py
index b2d6514b..69d53e9d 100644
--- a/invesalius/reader/dicom_reader.py
+++ b/invesalius/reader/dicom_reader.py
@@ -25,8 +25,8 @@ from multiprocessing import cpu_count
 import gdcm
 
 # Not showing GDCM warning and debug messages
-gdcm.Trace_DebugOff()
-gdcm.Trace_WarningOff()
+gdcm.Trace.DebugOff()
+gdcm.Trace.WarningOff()
 
 import glob
 import plistlib

Using python 3.10.9

It's user error on my end, attempting to create an ebuild for gentoo.

I want to convert the invesalius3 python script to execute program on windows by using pyinstaller. But the compiled exe always flashed and crashed. Maybe you can provide the configure file to convert python to exe.

Hi I am unable to connect the Fastrack tracker. When I try it prompts me to choose the Com Port, but there it displays me no options and says the tracker cannot connect. Is there a fix for this . I am using the latest release of the software.

The loading of the resonance happens in a strange way, I attach screen. The software does not allow me to select fiducials and markers on the 3D surface. And if I select them on the sections sometimes the point doesn't match the right position. Also I cannot run it with python or python3 I have to run it necessarily with pythonw, why?