
I am working through the docs of skimage and tried out the image conversions. The docs say that these functions *properly rescale* the values, but I'm not sure what is meant with *properly? * *IF *rescaling is meant to exploit the defined range between 0 and 1 for floats, that does not seem to happen: In [13]: data=np.uint(data) In [14]: data.max() Out[14]: 27 In [15]: data.min() Out[15]: 17 In [16]: fimg = skimage.img_as skimage.img_as_float skimage.img_as_int skimage.img_as_ubyte skimage.img_as_uint In [16]: fimg = skimage.img_as_float(data) In [17]: fimg.max() Out[17]: 6.2864273801181529e-09 In [18]: fimg.min() Out[18]: 3.9581209430373555e-09 In [19]: imshow(fimg) Out[19]: <matplotlib.image.AxesImage at 0xe38b1f0> In [20]: data.dtype Out[20]: dtype('uint32') Or do I misunderstand something? Best regards, Michael