summaryrefslogtreecommitdiff
path: root/README.md
blob: b0b1a45da3b8b636ec3607a237ed3d2b882e9150 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# OpenCV vs Flux.jl vs ImageFiltering.jl vs scipy vs pytorch

## Julia (ImageFiltering, Flux)
- precompile times are long (minutes)
- i find the syntax much nicer. Example: streaming operator (`.|>`): apply the function `Gray` to each value from `load("./cameraman.png")`
```julia
img_gray = load("./cameraman.png") .|> Gray
```
- i can interact with the running program (REPL)

## Python (scipy, pytorch)
- i find the syntax ugly..
- i can interact with the running program (REPL)

## OpenCV
- i find it a bit cumbersome, but workable
- compilation is relatively fast, but i can't interact with the running program

## Neural Network Libs (Flux.jl, Pytorch)
- they are heavier (might not be needed here, unless we do deep learning)