![]() ImageSeriesReader (0000021B082D3360): Non uniform sampling or missing slices detected, maximum nonuniformity:7.39539e-07 Output: WARNING: In d:\a\1\sitk-build\itk-prefix\include\itk-5.2\itkImageSeriesReader.hxx, line 480 ![]() Volume.point_data = data.flatten(order="F") Ps = pstats.Stats(pr, stream=s).sort_stats(sortby)ĭicom_files = dicom_reader.GetGDCMSeriesFileNames(folder)ĭata = (data // 256).astype(np.uint8) # Cast 16-bit to 8-bit Pv.rcParams = "fixed_point" # Windowsįolder = "C:\\path\\to\\DICOM\\stack\\folder" ![]() OS: Windows 10 Professional 圆4-bit, Build 1909ĭisplay GPU: 1x NVIDIA Quadro RTX4000 Software I've used cProfile to help identify problem areas (please see below).ĭICOM Image Dimensions: 2402 x 1301 pixels System / Hardware ParaView 5.9.1 (installed pre-built binary): ~24 seconds ParaView is My Code: ~13 minutes, 9 seconds Is there some way I can make this faster? Plotting volumes in pyvista is much slower than in ParaView. I use fixed_point since I am on Windows.) The results are the same regardless if I use the fixed_point vs. However, below you can see pyvista orients my data along the z-axis, when in fact it should be along the x-axis, and that it is garbled. ( NOTE: I cannot share the actual data because it is confidential. ParaView on the other hand plots things properly. Pyvista does not plot my volume along the correct axis and the output is garbled. After some analysis, it doesn't seem Qt threading is the culprit, so corresponding Qt code was removed. Update 1: Simplified problem even further to a smaller MRE. ![]()
0 Comments
Leave a Reply. |