Split the L0Smooth algorithm for Gray images and RGB images

[JKay0327]

In this work, there are two main mission:

• First, split the L0Smooth algorithm for Gray images and RGB images. smooth API no longer do pre-process for the input image. The processing works should be done in the algorithm part.
• Second, L0Smooth now support the widely-used OffsetArrays.

It's glad to see that the performance for processing Gray input improved a lot.

The performance after using ImageBase.fdiff:

julia> img_gray = testimage("cameraman");

julia> f = L0Smooth()
L0Smooth(0.02, 2.0, 100000.0)

julia> @btime smooth(img_gray, f)
  393.819 ms (1449 allocations: 138.77 MiB)

Current performance after split the L0Smooth algorithm:

julia> img_gray = testimage("cameraman");

julia> f = L0Smooth()
L0Smooth(0.02, 2.0, 100000.0)

julia> @btime smooth(img_gray , f)
  229.504 ms (1434 allocations: 130.94 MiB)

[codecov[bot]]

Codecov Report

Merging #7 (7b7ed21) into master (9a0354e) will not change coverage. The diff coverage is 100.00%.

@@            Coverage Diff            @@
##            master        #7   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            2         2           
  Lines           69       107   +38     
=========================================
+ Hits            69       107   +38     

Impacted Files	Coverage Δ
src/SmoothAPI/smooth.jl	100.00% <100.00%> (ø)
src/algorithms/l0_smooth.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 9a0354e...7b7ed21. Read the comment docs.