医学图像预处理----ManualNormalization（手动归一化，即窗宽窗位的设定）

通过给定的最大和最小窗口级别映射强度来标准化图像

sitk.IntensityWindowingImageFilter()将线性变换应用于用户定义的间隔内的输入图像的强度级别。 低于此间隔的值被映射到一个常数，区间内的值被映射到另一个常量

class ManualNormalization(object):
	"""
	Normalize an image by mapping intensity with given max and min window level
	"""

	def __init__(self,windowMin, windowMax):
		self.name = 'ManualNormalization'
		assert isinstance(windowMax, (int,float))
		assert isinstance(windowMin, (int,float))
		self.windowMax = float(windowMax)
		self.windowMin = float(windowMin)

	def __call__(self, sample):
		image, label = sample['image'], sample['label']
		intensityWindowingFilter = sitk.IntensityWindowingImageFilter()
		intensityWindowingFilter.SetOutputMaximum(255)
		intensityWindowingFilter.SetOutputMinimum(0)
		intensityWindowingFilter.SetWindowMaximum(self.windowMax);
		intensityWindowingFilter.SetWindowMinimum(self.windowMin);

		for channel in range(len(image)):
			image[channel] = intensityWindowingFilter.Execute(image[channel])

		return {'image': image, 'label': label}