System information (version)

• OpenCV => 4.5

Detailed description

The documentation for color conversions does not explain what _FULL does, e.g. https://docs.opencv.org/4.5.0/d8/d01/group__imgproc__color__conversions.html#gga4e0972be5de079fed4e3a10e24ef5ef0ae0070d7e97f1e565cc0992d038e5498e
I believe this should be explained for each color conversion as there are different inter

I'm receiving error with JPG (uppercase of jpg) image extension. This extension should be supported right?

ERROR Error resizing: Error: Expected one of: heic, heif, jpeg, jpg, png, raw, tiff, webp for format but received JPG of type string
at Object.invalidParameterError (/var/task/node_modules/sharp/lib/is.js:101:10)
at Sharp.toFormat (/var/task/node_modules/sharp/lib/output.js:168:

There is a set of Pixel Level transforms that is used in the work Benchmarking Neural Network Robustness to Common Corruptions and Perturbations

The authors also share the code => we can absorb some transforms that they have into the library.

https://github.com/hendrycks/robustness/blob/master/ImageNet-C/create_c/make_imagenet_c.py

• ShotNoise
• Defocus
• GlassBlur

A follow up on SixLabors/ImageSharp#1378 (comment).

Currently 32 bit test execution is only done for .NET Framework, with dotnet xunit which is an obsolete tool today, we need to adapt dotnet test, and add 32 bit CI targets for both net5.0 and netcoreapp3.1. Opening an issue to remember and track this debt.

See:
https://twitter.com/tannergooding/s

Description

I find myself repeating the following lines very often:

coords = ... # something that gives me an (npoints, ndim) array
indices = tuple(np.transpose(np.round(coords).astype(int)))
labels = np.zeros_like(image, dtype=np.uint32)
labels[indices] = np.arange(1, len(coords) + 1)

This is required for example to take a set of centroids and turn them into markers

In kornia.color.rgb_to_hsv, the computation of s is leading to nan value due to division by 0 when variable v contains 0.
s: torch.Tensor = deltac / v # saturation
I don t know what is the best solution to solve this. On my local machine, I add a small eps=1e-7 as a workaround:
s: torch.Tensor = deltac / (v + 1e-7) # saturation