Precompute lookup for N0f8

[timholy]

transform_density! is the main bottleneck for histogram equalization. When the number of possible values in the image is modest, it makes sense to precompute the entire lookup table and then apply it.

[codecov[bot]]

Codecov Report

Merging #55 (114d9b9) into master (d545b4e) will increase coverage by 0.04%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #55      +/-   ##
==========================================
+ Coverage   95.89%   95.94%   +0.04%     
==========================================
  Files          12       12              
  Lines         585      592       +7     
==========================================
+ Hits          561      568       +7     
  Misses         24       24              

Impacted Files	Coverage Δ
src/algorithms/common.jl	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update d545b4e...114d9b9. Read the comment docs.

[timholy]

Performancewise, this eliminates the gap between us and OpenCV in the "histeq" benchmark of https://github.com/JuliaImages/image_benchmarks, except at the two largest image sizes (where I suspect OpenCV switches to a multithreaded algorithm).

[timholy]

Maybe get a release soon? I'm not sure whether you want to merge some of the other changes first.

[zygmuntszpak]

I was initially planning to incorporate the PiecewiseLinearStretching stuff, but may as well release your performance improvements first.

[zygmuntszpak]

@JuliaRegistrator register

[JuliaRegistrator]

Comments on pull requests will not trigger Registrator, as it is disabled. Please try commenting on a commit or issue.