gidesa 22 Posted April 1, 2023 Hello, is ready to download and test the wrapper around the Opencv version 4.6 C++ API . This is the more recent Opencv stable version. The old pure C API misses all the functions implemented only as C++ classes, starting from the Machine Language module (ML), that was present since the Opencv v 1.0 . Note: from v 2.4 onwards the old C API is always present, although no more officially supported. This version of Opencv, among many other things, has a Deep Neural Network (DNN) module, capable of load and use the most common neural net formats: Tensorflow, Torch, Caffe, Darknet, ONNX. The wrapper supports this, of course, and has many improvements compared to my precedent wrapper for Opencv v 2.4.31 . All suggestions are welcome. On Github: https://github.com/gidesa/ocvWrapper46 Thanks Giando 7 1 Share this post Link to post
gidesa 22 Posted April 7, 2023 Hello to all, I have released a new version of wrapper library, with many improvements, and instructions for build on Linux, too. https://github.com/gidesa/ocvWrapper46 3 Share this post Link to post
gidesa 22 Posted May 4, 2023 Hello again, the OpencvWrapper46 version 1.1 is released and ready to test. There are four new VCL components and other classes that great simplify the use of Opencv, especially for object recognition, and face recognition/detection. https://github.com/gidesa/ocvWrapper46 Thanks Giando Share this post Link to post
shineworld 73 Posted May 4, 2023 I was not able to run pre-made samples. They are 32-bit not 64-bit and library README.MD say that only 64-bit are possibile. I've tried to activate 64-bit platform in samples project but only 32-bit it is possible. Share this post Link to post
gidesa 22 Posted May 4, 2023 Hello, you are right, all Delphi examples are compiled as 32 bit. I develop on a Delphi 32 bit version. If not possible from Delphi IDE, you could activate 64 bit platform by hand modify the .DPROJ file: in platform group add a line for Win64 <Platforms> <Platform value="Win32">True</Platform> <Platform value="Win64">True</Platform> </Platforms> Official version of Opencv has only the 64 bit binaries. You have to compile Opencv in 32 bit by yourself, has I have done. Anyway, I could download to github the 32 bit versions of Opencv DLL. Write me if this could be useful for you. Share this post Link to post
shineworld 73 Posted May 5, 2023 13 hours ago, gidesa said: Anyway, I could download to github the 32 bit versions of Opencv DLL. This can help because I was able to find and download ONLY OpenCV DLL for 64-bits. Share this post Link to post
gidesa 22 Posted May 5, 2023 Hello, in github opencv4.6-x86 directory now there are the 32 bit binaries. Regards Share this post Link to post
marcovth 0 Posted June 11, 2023 (edited) Is there any chance this OpenCV wrapper could work for Android apps? I am exploring if I could make a "Tourist filter" app with life view, similar to the Spectre app for iOS. https://beltoforion.de/en/the-tourist-filter/ Edited June 11, 2023 by marcovth Share this post Link to post
gidesa 22 Posted June 21, 2023 Hello marcovth, I think that in principle the wrapper could be compiled as shared library also under Android, using the Linux examples as a starting point. But I have no experience on Android, sorry. Relevant sites, for example: https://blog.arik.io/compiling-c-code-as-android-shared-libraries-6924803c00db https://stackoverflow.com/questions/70053057/creating-native-shared-library-with-android-studio Share this post Link to post
gidesa 22 Posted March 13 Hello, the Opencv 4.6 wrapper repository now has new examples, better Pascal classes, much better documentation. And instructions for compile on MacOS, too. Enjoy! Github repo https://github.com/gidesa/ocvWrapper46/ 1 2 Share this post Link to post
gidesa 22 Posted November 14 Hello again, released many improvements and new examples: - classes that encapsulate OCV vectors and pointers. - a component to write movies from a video stream. - an example for movie writing, - an example that uses Machine Learning classifiers to recognize handwritten digits. - a detailed documentation See at https://github.com/gidesa/ocvWrapper46 1 Share this post Link to post