ImageCore.jl
ImageCore.jl copied to clipboard
JuliaImages
Julia types for representing images
MWE: ```julia (jl_TQljLn) pkg> st Status `/tmp/jl_TQljLn/Project.toml` [a09fc81d] ImageCore v0.9.4 julia> x = rand(3,100,100,10); # 10 images julia> @views x1 = x[:,:,:,1]; # first image julia> size(x1) (3, 100, 100)...
We used to have one single package `ImageCore` that keeps all the common small functions so that our other toolboxes can reuse them. But it's a growing need to also...
It's unclear why this package is added to ImageCore so maybe we should clean up this dependency. I don't plan to merge this until our next breaking release, but it's...
Due to the same spirit to https://github.com/JuliaGraphics/ColorTypes.jl/pull/270 but I'm not very sure of it. Probably we'd like to limit ourselves to a smaller set of precompilation statements? ```julia julia> @time...
Is there any reason why we use `mappedarray` rather than plain `reinterpret` here? https://github.com/JuliaImages/ImageCore.jl/blob/edd20781c875d77d03531d3592e812736e35def2/src/ImageCore.jl#L124 This adds a few performance overheads to ImageMorphology https://github.com/JuliaImages/ImageMorphology.jl/issues/103 because `pointer` of MappedArray is an invalid...
Four packages are reexported by ImageCore: - IndirectArrays - MappedArrays - OffsetArrays - TiledIteration The extra time overhead seems good to me. (Although in static compiled languages like C++ there's...
The 0-to-1 colorscale does not have meaning for colors like HSV, so supporting arbitrary colorants seems wrong. I noticed this while working on https://github.com/JuliaImages/FluorophoreColors.jl/pull/4. We might not want to merge...
This provides a more convenient MATLAB interface compared to the composition of `PermutedDimsArray`, `colorview`, etc (CRef #178) Although our naming guidelines recommend not adding `im_` prefix, I feel `from_matlab` is...
This is a re-post from a recent slack discussion > > I don't think JuliaImages will try to break the "an image is `AbstractArray{
This does not fix #174 but could be potentially a good change by eliminating one mapped array wrapper.