This package is devoted for detection of shape in binarized images. We highlight that this package is BETA QUALITY, just for academic purposes.
All implementations were made using Julia Language and in this stage the code can be very rough. There is three algorithms implemented:
- one based on Gauss-Newton Method
- the HT for circle detection
- a new hybrid algorithm based on Gauss-Newton and HT.
For test all algorithm you need
- Install Julia Language
- Install some dependencies
For install all dependencies just type in Julia REPL:
Pkg.add("Images")Pkg.add("Colors")Pkg.add("FixedPointNumbers")Pkg.add("ImageView")
After that, you need to include the setup file. For this, just type
include("setup_tests.jl").
In order to run some problem type something like:
run_prob("datafile-50-300-30-10","HT",20,true,false), for artificial problems or
run_prob("moedas","HT",20,true,true), for real problems. Change HT by GN or HTGN in order to test other methods.
All images (real or artificial) are in /real_examples folder or /synthetic_example folder.
The folder /scripts contain some tests in artificial instances.
This implementation was made by:
- Emerson Vitor Castelani
- Wesley V. I. Shirabayashi
- Jair da Silva
- Eduardo Neves