I used the following flags and cmake to configure (followed by ninja to build ITK)
I got an error at HDF5 build step. So, I followed section 2 in this link to fix that error. I seemed to get past that stage but i now get the following error:
ninja: build stopped: command failed
The lines above this error are the following:
InsightToolkit-4.9.1/Modules/Core/Common/include/itkPoint.h:308:48: note: previous declaration is here typedef typename PointContainerType::Element PointType;
3 warnings generated.
This is my first attempt at using ninja and emscripten. I might be posting an incomplete error message above. In CMake, I would look at CMakeError.log but i do not know the equivalent for that file in ninja. Can someone please point me in the right direction and help me build ITK using emscripten?
Thanks @dzenanz. I did not try 5.1 but I got errors related to threading in ITK-5.0.1. The errors were in itkMultiThreaderBase.cxx about undeclared variable m_GlobalMaximumNumberOfThreads. I changed the corresponding header file and created a variable for m_GlobalMaximumNumberOfThreads but that just led to more errors. Did I do something wrong? If not, have they been fixed in ITK 5.1 for clang compiler?
Thanks again, @matt.mccormick. The example link you provided me helped me a ton and i’m up and running.
Is there a way to pass an Arraybuffer (preferably as 16-bit and not 8-bit) to ITK based C++ code. In the example you provided, the input image to C++ is read from the disk. In my application, this file is read on the browser side using UTIF and I have access just to the Arraybuffer from UTIF’s reader. Do you have any examples for extracting pixel intensities from that arraybuffer and passing it to the C++ code?
Thanks @matt.mccormick. arrayBuffer that is passed to me has been encoded. Below is the code that the front-end developer on my team (working on HTML side) gave me.
// Create a new XMLHttpRequest object to read the mock TIFF image as ArrayBuffer
const xhr = new XMLHttpRequest();
// Configure it: GET-request for the URL
xhr.open('GET', 'assets/tif/test.tif', true); // read test.tif bundled in the assets folder
xhr.responseType = 'arraybuffer'; // get array buffer
xhr.timeout = 10000; // timeout in ms, 10 seconds
In the example code, front-end developer is creating a HTTP request to read a test.tif I gave him. test.tif is a 300x300 scalar image with unsigned char as its pixel type. Therefore, I expected arrayBuffer variable in the last line of the code block above to be a 90,000x1 array. However, when I print the contents of arrayBuffer to a text file it has additional 225 elements. And, it does not have the tiff tags associated with it. It looks like an encoded array. Is there a simple way to decode this array?
I tried to copy itk-js’ appoach for readImageHTTP and use axios reader do but realized that the input file has to be a json with a 90,000x1 array for pixel intensities. I seem to be stuck in a loop with this approach.
'am sure i’m missing something really simple. Can you please point me in the right direction?
Thanks again @matt.mccormick. I looked at source code for readImageArrayBuffer and it looked like even it was expecting a decoded tiff. I could be wrong here but my front end team figured out a way to decode TIFFs and send me just the pixel intensities. I’m using that in my C++ code to create an ITK image and set up my C++ processing pipeline.