New release IPCV 4.5.0.1

[mottelet]

Hello,

Some recent bug fixes which were fixed in the master branch have been ported to the 4.5 branch of IPCV (based on OpenCV). Apart from the build of bundled OpenCV which has been improved, the new feature is the possibility of reading png/tiff images while keeping the alpha channel, e.g.

im = imread(fullpath(getIPCVpath()+"/images/baboon.png"));
imshow(im);
im = imread(fullpath(getIPCVpath()+"/images/puffin.png"),IMREAD_UNCHANGED=1);
imshow(im);

The new release has been numbered 4.5.0.1, and until we update the bundled OpenCV library, subsequent versions will be numbered 4.5.0.n.

Enjoy !

S.

[stefa24]

Hi all
I am running with scilab 2025 0 0
I am trying to instal IPCV but after more than 30 min nothing happen, the installation is still ongoing after more than 30min

I followed what provided here

https://scilab.discourse.group/t/unable-to-install-ipcv/739

and removed all configuration files

C:\Users\stefano\AppData\Roaming\Scilab\scilab-2025.0.0

Version 4.5.0.1AuthorTan Chin LuhMaintainers Stéphane MOTTELET Chin Luh Tan Vincent COUVERTCategoryImage ProcessingWebsite https://github.com/tanchinluh/IPCVLicenseGPL (3.0)Creation DateFebruary 25, 2025 Source created onScilab 2025.0.xBinaries available on
Scilab 2025.0.x:

thanks for your help

stefano

[mottelet]

Hello,

Please don’t cross post to multiple topics.

The IPCV toolbox is quite heavy because of the bundled OpenCV library. Did you check that your internet connection is working fine ?

S.

[stefa24]

hi
thanks and apologize
internet is working fine for what i can undestand
i am quite new with scilab
can I installl manually and how from the .zip file?

[IPCV-4.5.0.1.bin.2025.0.x64.Windows.zip ]

Windows 64-bit binary for Scilab 2025.0.x

[IPCV-4.5.0.1.zip ]

Source code archive

stefano

[mottelet]

Hi again,

Supposing that you have changed the current directory to the location where the file is actually present:

atomsSetConfig offLine true
atomsInstall IPCV-4.5.0.1.bin.2025.0.x64.Windows.zip
atomsLoad IPCV

S.

[stefa24]

many thanks

it is working

stefano

[P3770K]

Hi ! I tried to install ICPV module on my laptop but I still get error message:

atomsInstall: Error while building the toolbox ‘C:\Users\Blabla\AppData\Roaming\Scilab\scilab-2025.1.0\atoms\x64\IPCV\4.5.0.2’.
à la ligne 52 de la fonction atomsError ( C:\Program Files\scilab-2025.1.0\modules\atoms\macros\atoms_internals\atomsError.sci ligne 66 )
à la ligne 418 de la fonction atomsInstall ( C:\Program Files\scilab-2025.1.0\modules\atoms\macros\atomsInstall.sci ligne 434 )

Other modules got installed fine…

[mottelet]

Hello,

You are using Scilab 2025.1, so the latest available version of IPCV for this version of Scilab is IPCV 4.5.0.1. So, if you want to stick to Scilab 2025.1 you have to specify the version,

atomsInstall(["IPCV","4.5.0.1"])

otherwise atoms will try to install IPCV 4.5.0.2, the latest version of IPCV, which is available only for Scilab 2026. The actual mecanism of atoms is not clever at all, as it allows to try to install under Scilab 2025 a module built for Scilab 2026…
However I advise you to install Scilab 2026 and install the latest IPCV version, it will be as simple as

atomsInstall IPCV

S.

[P3770K]

OK so :

• I tried this (on Scilab 2025)

  • atomsInstall([“IPCV”,“4.5.0.1”])

• It didn’t work

• I tried to install Scilab 2026 and IPCV ==> seemed to work but 1) lot of other modules (distfun, GUIbuilder…) failed to install and 2) after launch, atoms seemed unable to locate and launch the modules… (and if I remember well it’s a pita to find/change atoms path…)

• I came back to Scilab 2025 and saw an error about loader.sce missing in “C:\Users\Blabla\AppData\Roaming\Scilab\scilab-2025.1.0\atoms\x64\guibuilder\4.2.3\”

• I found loader.sce in a neighbour file (C:\Users\Blabla\AppData\Roaming\Scilab\scilab-2025.1.0\atoms\x64\guibuilder\guibuilder-4.2.3\) and copy/pasted it…

• IPCV is working (on Scilab 2025) !

I don’t get how repairing a missing file in GUIbuilder influenced IPCV but I’m just here to share…

Thanks for your response @mottelet at least I found a workaround…