ITK 连通域分析

连通域

一个区域D中，任一闭曲线所围区域，都属于区域D，称为“单连通区域”。

连通域形状分析

1. 创建二值图像

1 using ImageType = itk::Image<unsigned char, 2>; 2 3 ImageType::Pointer CreateMaskImage() 4 43 44 return image; 45 }

生成图像如下

2. 打印所有连通区域信息

1 typedef unsigned long LabelType; 2 typedef itk::ShapeLabelObject< LabelType, 2 > LabelObjectType; 3 typedef itk::LabelMap< LabelObjectType > LabelMapType; 4 typedef itk::BinaryImageToShapeLabelMapFilter< ImageType, LabelMapType > I2LType; 5 I2LType::Pointer i2l = I2LType::New(); 6 i2l>SetInput(mask); 7 i2l>SetFullyConnected(false); 8 i2l>SetInputForegroundValue(1); 9 i2l>SetOutputBackgroundValue(0); 10 i2l>Update(); 11 12 LabelMapType::Pointer labelMap = i2l>GetOutput(); 13 for (unsigned int label = 1; label <= labelMap>GetNumberOfLabelObjects(); label++) 14

打印信息如下：

1 ShapeLabelObject (000002469EAA6CE0) 2 RTTI typeinfo: class itk::ShapeLabelObject<unsigned long,2> 3 Reference Count: 1 4 LineContainer: 000002469EAA6CF0 5 Label: 1 6 NumberOfPixels: 616 7 PhysicalSize: 616 8 Perimeter: 127.254 9 NumberOfPixelsOnBorder: 0 10 PerimeterOnBorder: 0 11 PerimeterOnBorderRatio: 0 12 Elongation: 1 13 Flatness: 1 14 Roundness: 0.691393 15 Centroid: [50, 50] 16 BoundingBox: ImageRegion (000002469EAA6D20) 17 Dimension: 2 18 Index: [36, 36] 19 Size: [29, 29] 20 EquivalentSphericalRadius: 14.0028 21 EquivalentSphericalPerimeter: 87.9823 22 EquivalentEllipsoidDiameter: [28.0056, 28.0056] 23 PrincipalMents: [61.9156, 61.9156] 24 PrincipalAxes: 25 1 0 26 0 1 27 FeretDiameter: 0 28 m_OrientedBoundingBoxSize: [0, 0] 29 m_OrientedBoundingBoxOrigin: [0, 0] 30 31 ShapeLabelObject (000002469EAA7030) 32 RTTI typeinfo: class itk::ShapeLabelObject<unsigned long,2> 33 Reference Count: 1 34 LineContainer: 000002469EAA7040 35 Label: 2 36 NumberOfPixels: 145 37 PhysicalSize: 145 38 Perimeter: 41.524 39 NumberOfPixelsOnBorder: 0 40 PerimeterOnBorder: 0 41 PerimeterOnBorderRatio: 0 42 Elongation: 1 43 Flatness: 1 44 Roundness: 1.02799 45 Centroid: [70, 80] 46 BoundingBox: ImageRegion (000002469EAA7070) 47 Dimension: 2 48 Index: [64, 74] 49 Size: [13, 13] 50 EquivalentSphericalRadius: 6.79374 51 EquivalentSphericalPerimeter: 42.68 52 EquivalentEllipsoidDiameter: [13.5875, 13.5875] 53 PrincipalMents: [11.5172, 11.5172] 54 PrincipalAxes: 55 1 0 56 0 1 57 FeretDiameter: 0 58 m_OrientedBoundingBoxSize: [0, 0] 59 m_OrientedBoundingBoxOrigin: [0, 0] 60 61 ShapeLabelObject (000002469EAA7170) 62 RTTI typeinfo: class itk::ShapeLabelObject<unsigned long,2> Reference Count: 1 64 LineContainer: 000002469EAA7180 65 Label: 3 66 NumberOfPixels: 231 67 PhysicalSize: 231 68 Perimeter: 59.5651 69 NumberOfPixelsOnBorder: 0 70 PerimeterOnBorder: 0 71 PerimeterOnBorderRatio: 0 72 Elongation: 1.91485 73 Flatness: 1.91485 74 Roundness: 0.904522 75 Centroid: [110, 105] 76 BoundingBox: ImageRegion (000002469EAA71B0) 77 Dimension: 2 78 Index: [100, 100] 79 Size: [21, 11] 80 EquivalentSphericalRadius: 8.57494 81 EquivalentSphericalPerimeter: 53.8779 82 EquivalentEllipsoidDiameter: [12.3935, 23.7317] 83 PrincipalMents: [10, 36.6667] 84 PrincipalAxes: 85 0 1 86 1 0 87 FeretDiameter: 0 88 m_OrientedBoundingBoxSize: [0, 0] 89 m_OrientedBoundingBoxOrigin: [0, 0] 90

3. 统计灰度信息

1 typedef itk::BinaryImageToStatisticsLabelMapFilter< ImageType, ImageType > ConverterType; 2 ConverterType::Pointer converter = ConverterType::New(); 3 converter>SetInput(reader>GetOutput()); 4 converter>SetFeatureImage(reader2>GetOutput()); 5 converter>SetInputForegroundValue(1); 6 converter>SetFullyConnected(false); 7 converter>Update(); 8 converter>GetOutput()>PrintLabelObjects();

4. 过滤连通区域

1 typedef unsigned long LabelType; 2 typedef itk::ShapeLabelObject< LabelType, 2 > LabelObjectType; 3 typedef itk::LabelMap< LabelObjectType > LabelMapType; 4 typedef itk::BinaryImageToShapeLabelMapFilter< ImageType, LabelMapType > I2LType; 5 I2LType::Pointer i2l = I2LType::New(); 6 i2l>SetInput(mask); 7 i2l>SetFullyConnected(false); 8 i2l>SetInputForegroundValue(1); 9 i2l>SetOutputBackgroundValue(0); 10 i2l>Update(); 11 12 LabelMapType::Pointer labelMap = i2l>GetOutput(); 13 for (unsigned int label = 1; label <= labelMap>GetNumberOfLabelObjects(); label++) 14 19 20 std::vector< LabelObjectType::Pointer > filtered_label_objects; 21 for (LabelMapType::Iterator it(labelMap); !it.IsAtEnd(); ++it) 22 27 28 for (int i = 0; i < filtered_label_objects.size(); i++) 29 32 33 typedef itk::LabelMapToBinaryImageFilter< LabelMapType, ImageType> L2IType; 34 L2IType::Pointer l2i = L2IType::New(); 35 l2i>SetInput(labelMap); 36 l2i>Update(); 37 38 using WriterType = itk::ImageFileWriter<ImageType>; 39 WriterType::Pointer writer = WriterType::New(); 40 writer>SetFileName("mask.mha"); 41 writer>SetInput(l2i>GetOutput()); 42 writer>SetImageIO(itk::MetaImageIO::New()); 43 writer>Update();

效果如下：

总结
通过连通域形状分析可以过滤到阈值分割中的错误部分，另外ITK还提供了itk::BinaryImageToStatisticsLabelMapFilter类，可以提供原图区域的统计信息。

itk::ShapeLabelObject提供的形状信息包括：

1. NumberOfPixels 像素数
2. PhysicalSize 物理尺寸
3. Perimeter 周长
4. NumberOfPixelsOnBorder 边界像素数
5. PerimeterOnBorder 边界周长
6. PerimeterOnBorderRatio 边界周长比
7. Elongation 伸长率
8. Flatness 平整度
9. Roundness 圆度
10. Centroid 质心
11. BoundingBox 包围盒
12. EquivalentSphericalRadius 等效球面半径
13. EquivalentSphericalPerimeter 等效球面周长
14. EquivalentEllipsoidDiameter 等效椭球直径
15. PrincipalMents 主力矩
16. PrincipalAxes 主轴
17. FeretDiameter

itk::StatisticsLabelObject提供的统计信息包括：

1. Minimum
2. Maximum
3. Mean
4. Sum
5. StandardDeviation 标准差
6. Variance 方差
7. Median
8. Skewness 偏态
9. Kurtosis 峰度
10. WeightedElongation 加权伸长率
11. WeightedFlatness 加权平面度
12. MaximumIndex
13. MinimumIndex
14. CenterOfGravity 重心
15. WeightedPrincipalMents 加权主力矩
16. WeightedPrincipalAxes 加权主轴

上一篇：SimpleITK 获取二值图像bbox

下一篇：VTK 问题整理

---

ITK