pyradiomics特征提取过程中报错

安装好pyradiomics可以正常import之后，尝试用示例数据运行一下流程，在提取特征这一步出了问题。

result = extractor.execute(imagePath, maskPath)

运行之后显示报错

result = extractor.execute(imagePath, maskPath)
Traceback (most recent call last):

  File "C:\Users\LeiLei\AppData\Local\Temp\ipykernel_15372\454753011.py", line 1, in <module>
    result = extractor.execute(imagePath, maskPath)

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\featureextractor.py", line 325, in execute
    featureVector.update(self.computeFeatures(inputImage, inputMask, imageTypeName, **inputKwargs))

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\featureextractor.py", line 501, in computeFeatures
    for (featureName, featureValue) in six.iteritems(featureClass.execute()):

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\base.py", line 185, in execute
    self._calculateSegment()

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\base.py", line 227, in _calculateSegment
    for success, featureName, featureValue in self._calculateFeatures():

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\base.py", line 234, in _calculateFeatures
    self._initCalculation(voxelCoordinates)

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\glcm.py", line 111, in _initCalculation
    self.P_glcm = self._calculateMatrix(voxelCoordinates)

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\radiomics\glcm.py", line 148, in _calculateMatrix
    P_glcm = numpy.delete(P_glcm, emptyGrayLevels - 1, 1)

  File "<__array_function__ internals>", line 6, in delete

  File "C:\Users\LeiLei\anaconda3\envs\radio\lib\site-packages\numpy\lib\function_base.py", line 4552, in delete
    keep[obj,] = False

IndexError: arrays used as indices must be of integer (or boolean) type

参照github的问答（https://github.com/AIM-Harvard/pyradiomics/issues/592），把imageoperations.py里的代码改了一下，在第157行增加了一小段。

原来：
  return parameterMatrix, binEdges
现在：
  return parameterMatrix.astype('int'), binEdges

问题没有解决，继续找，看到有人说修改glcm.py文件。在该文件的148行增加一条语句：

    if list(emptyGrayLevels):

变成：

    if list(emptyGrayLevels):
      P_glcm = numpy.delete(P_glcm, emptyGrayLevels - 1, 1)
      P_glcm = numpy.delete(P_glcm, emptyGrayLevels - 1, 2)

还是报错
然后我就把pyradiomics移除重新安装了一下，还是不行。现在开始怀疑是不是包的问题。
但是我换了示例数据run一下，就没问题了。
所以问题还是出在示例数据上。Brain1这个示例数据不对劲。
这个故事告诉我们，不要迷信示例数据。弹报错的时候先检查数据本身可能是更便捷的思路。