Skip to content

HolyLab/RegisterMismatch.jl

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

78 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RegisterMismatch.jl

CI Coverage Stable docs Dev docs

FFT-based image mismatch computation for translation-based image registration. This package is the CPU backend in the HolyLab image registration stack; a GPU counterpart is provided by RegisterMismatchCuda.jl.

Installation

This package is registered in the HolyLabRegistry. Add the registry once, then install as usual:

using Pkg
pkg"registry add General https://github.com/HolyLab/HolyLabRegistry.git"
Pkg.add("RegisterMismatch")

Quick start

using RegisterMismatch

# Two images, moving shifted by (2, 3) relative to fixed
fixed  = zeros(Float32, 32, 32); fixed[10:22, 10:22] .= 1f0
moving = circshift(fixed, (2, 3))

# Find the optimal integer-pixel shift
shift = register_translate(fixed, moving, (5, 5))
# CartesianIndex(2, 3)

For non-uniform or large displacements, use mismatch_apertures to obtain a grid of local mismatches and recover a smooth displacement field.

See the documentation for a full guide, including preprocessing (highpass, nanpad), aperture-based registration, bias correction, and the low-level CMStorage workflow for repeated computations against the same fixed image.

About

This package is separated from BlockRegistraton.

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages