ashwani-rathee commented 3 years ago

[x] Supposed to provide demonstrations and examples for ImageBinarization.jl

Status : Early Stage, Not Ready

Preview Link :

Knowingly/unknowingly, current version does show http related methods and metaprogramming support usage here which might be a good thing.

Notes:

To binarize a brain scan(.nii)(3d image), needs to be figured out
ImageMagick.load loses a lot of detail this way of url download
```
using HTTP, ImageMagick
```

function get_image(url) response = HTTP.get(url) buffer = IOBuffer(response.body) img = ImageMagick.load(buffer) end

This returns a very bad image for some reason


Better method to download images
```julia
using Images
function get_image(url)
    img = mktemp() do fn, f
        download(url, fn)
        load(fn)
    end
    img_resized = imresize(img, ratio = 1 / 2)
end

img_coins = get_image("https://i.imgur.com/91IOKtJ.png")

ashwani-rathee commented 2 years ago

Actually, there isn't much to add with what I know. I think we can add the build histogram(with its graph) and show the threshold..But it would require me to use plots here, which might slow things more I think

johnnychen94 commented 2 years ago

I'm a little bit concerned about the current version because it adds basically nothing new compared to the existing index page of ImageBinarizations. The API itself is quite simple to understand and use.

I feel we need some application that uses ImageBinarization as a part of its pipeline. The sudoku solver sounds interesting, fingerprint enhancement together with ImageMorphology seems also a good candidate. How do you think?

JuliaImages / juliaimages.github.io

Image Binarization #201

This returns a very bad image for some reason