ITK 实例17 阈值水平集算法对脑部MHA文件进行三维分割

1 #include "itkImage.h" 2 #include "itkThresholdSegmentationLevelSetImageFilter.h" 3 4 #include "itkFastMarchingImageFilter.h" 5 #include "itkBinaryThresholdImageFilter.h" 6 #include "itkImageFileReader.h" 7 #include "itkImageFileWriter.h" 8 #include "itkZeroCrossingImageFilter.h" 9 10 11 int main( int argc, char *argv[] ) 12 */ 25 /*现在我们使用一个特殊的像素类型和维来定义图像类型。在这种情况下我们将使用二维 26 浮点型图像*/ 27 typedef float InternalPixelType; 28 const unsigned int Dimension = 3; 29 typedef itk::Image< InternalPixelType, Dimension > InternalImageType; 30 //接下来的几行使用New()方式实例化一个ThresholdSegmentationLevelSetImageFilter： 31 typedef unsigned char OutputPixelType; 32 typedef itk::Image< OutputPixelType, Dimension > OutputImageType; 33 typedef itk::BinaryThresholdImageFilter<InternalImageType, OutputImageType> 34 ThresholdingFilterType; 35 36 ThresholdingFilterType::Pointer thresholder = ThresholdingFilterType::New(); 37 38 thresholder>SetLowerThreshold( 1000.0 ); 39 thresholder>SetUpperThreshold( 0.0 ); 40 41 thresholder>SetOutsideValue( 0 ); 42 thresholder>SetInsideValue( 255 ); 43 44 typedef itk::ImageFileReader< InternalImageType > ReaderType; 45 typedef itk::ImageFileWriter< OutputImageType > WriterType; 46 47 ReaderType::Pointer reader = ReaderType::New(); 48 WriterType::Pointer writer = WriterType::New(); 49 //输入图像 50 reader>SetFileName( "BrainProtonDensity3Slices.mha" ); 51 //保存图像 52 writer>SetFileName( "naoshi_Threshold_LevelSet.mha" ); 53 54 typedef itk::FastMarchingImageFilter< InternalImageType, InternalImageType > 55 FastMarchingFilterType; 56 57 FastMarchingFilterType::Pointer fastMarching = FastMarchingFilterType::New(); 58 59 typedef itk::ThresholdSegmentationLevelSetImageFilter< InternalImageType, 60 InternalImageType > ThresholdSegmentationLevelSetImageFilterType; 61 ThresholdSegmentationLevelSetImageFilterType::Pointer thresholdSegmentation = 62 ThresholdSegmentationLevelSetImageFilterType::New(); /*对于ThresholdSegmentationLevelSetImageFilter，缩放比例参数用来平衡从等式（9 3） 64 中的传播(膨胀)和曲率(表面平滑)系数的影响。这个滤波器中不使用水平对流系数。使用 65 SetPropagationScaling()和SetCurvatureScaling()方式来设置这些系数。在这个例子中这两个 66 系数都设置为1.0*/ 67 thresholdSegmentation>SetPropagationScaling( 1.0 ); 68 /*if ( argc > 8 ) 69 72 else 73 */ 76 77 thresholdSegmentation>SetMaximumRMSError( 0.02 ); 78 thresholdSegmentation>SetNumberOfIterations( 1200 );//设置最大迭代次数 79 /* 收敛标准MaximumRMSError和MaximumIterations设置的和前面例子中的一样。现在我 80 们设置上下门限U和L，以及在初始化模型时使用的等值面值*/ 81 //上门限值（U） 82 thresholdSegmentation>SetUpperThreshold( ::atof("250") ); 83 //下门限值（L） 84 thresholdSegmentation>SetLowerThreshold( ::atof("210") ); 85 thresholdSegmentation>SetIsoSurfaceValue(0.0); 86 87 thresholdSegmentation>SetInput( fastMarching>GetOutput() ); 88 thresholdSegmentation>SetFeatureImage( reader>GetOutput() ); 89 thresholder>SetInput( thresholdSegmentation>GetOutput() ); 90 writer>SetInput( thresholder>GetOutput() ); 91 92 typedef FastMarchingFilterType::NodeContainer NodeContainer; 93 typedef FastMarchingFilterType::NodeType NodeType; 94 95 NodeContainer::Pointer seeds = NodeContainer::New(); 96 97 InternalImageType::IndexType seedPosition; 98 //设置种子点位置 99 seedPosition[0] = atoi("84"); 100 seedPosition[1] = atoi("126"); 101 seedPosition[2] = atoi("2"); 102 //表面到种子点的最初距离 103 const double initialDistance = atof("5");//5 104 105 NodeType node; 106 107 const double seedValue = initialDistance; 108 109 node.SetValue( seedValue ); 110 node.SetIndex( seedPosition ); 111 112 seeds>Initialize(); 113 seeds>InsertElement( 0, node ); 114 115 fastMarching>SetTrialPoints( seeds ); 116 117 fastMarching>SetSpeedConstant( 1.0 ); 118 /*调用writer上的Updata()方法引发了管道的运行。通常在出现错误和抛出异议时，从一 119 个try / catch模块调用updata：*/ 120 try 121 130 catch( itk::ExceptionObject & excep ) 131 136 137 std::cout << std::endl; 138 std::cout << "Max. no. iterations: " << thresholdSegmentation>GetNumberOfIterations() << std::endl; 139 std::cout << "Max. RMS error: " << thresholdSegmentation>GetMaximumRMSError() << std::endl; 140 std::cout << std::endl; 141 std::cout << "No. elpased iterations: " << thresholdSegmentation>GetElapsedIterations() << std::endl; 142 std::cout << "RMS change: " << thresholdSegmentation>GetRMSChange() << std::endl; 143 144 typedef itk::ImageFileWriter< InternalImageType > InternalWriterType; 145 //fastMarching快速步进输出水平集 146 InternalWriterType::Pointer mapWriter = InternalWriterType::New(); 147 mapWriter>SetInput( fastMarching>GetOutput() ); 148 mapWriter>SetFileName("fastMarchingImage.mha"); 149 mapWriter>Update(); 150 //阈值水平集分割的速度（传播系数P）图像 151 InternalWriterType::Pointer speedWriter = InternalWriterType::New(); 152 speedWriter>SetInput( thresholdSegmentation>GetSpeedImage() ); 153 speedWriter>SetFileName("speedTermImage.mha"); 154 speedWriter>Update(); 155 156 return EXIT_SUCCESS; 157 }

上一篇：OpenCV3.2图像分割 实例9：K-Means-证件照背景替换

下一篇：OpenCV3.2图像分割 实例10：绿幕视频背景替换

---

ITK