Missing clipping of dtype limits in is_low_contrast function

[jo-fra]

When using the skimage.exposure.is_low_contrast function I observe different results if

a. passing a grayscale image of type uint8 with pixel range [0, 255] that I converted beforehand

versus:

b. passing an RGB image and let it be converted within is_low_contrast by line:

scikit-image/skimage/exposure/exposure.py

Line 833 in 5e74a4a

image = rgb2gray(image)

The reason is that after conversion by rgb2gray the image is of type float64 and the pixel range [0.0, 1.0].
However, in the following line dtype_limits is called with clip_negative=False:

scikit-image/skimage/exposure/exposure.py

Line 835 in 5e74a4a

dlimits = dtype_limits(image, clip_negative=False)

Which results in dlimits=(-1,1) and an unexpected value of the calculated ratio.

I expect dtype_limits should be called with clip_negative=True to receive the expected ratio.

[stefanv]

Yeah, yikes, this is a bug, but also a subtle one because technically float dtypes do support negative numbers. @jni until skimage 2.0, would it make sense to change that range to (0, 1) everywhere instead of (-1, 1)?

[github-actions]

Hey, there hasn't been any activity on this issue for more than 180 days. For now, we have marked it as "dormant" until there is some new activity. You are welcome to reach out to people by mentioning them here or on our forum if you need more feedback! If you think that this issue is no longer relevant, you may close it by yourself; otherwise, we may do it at some point (either way, it will be done manually). In any case, thank you for your contributions so far!