Hi Christopher On 9 February 2016 at 11:23, Christopher <spagodo@gmail.com> wrote:
The reason I wanted to use img_as_float is because I eventually plan to use many functions and I thought I should get the image into a format that scikit-image likes before proceeding. However, as you suggested, the function I want to use immediately is transform.rotate() which does have a preserve scale option, and that did fix my problem. Should I just use this on every function?
From there on, you should have no problems. If, however, you need to
If you know the range of your input data, then you could rescale the image to be between -1 and 1: exposure.rescale_intensity(data, in_range=(-10, 10)) preserve your original scale, you can either use `preserve_range` or rescale your image afterwards. The latter approach is dangerous, though, since you don't know how the range of values get influenced by the operations you performed on your image. Regards Stéfan