foureg 0.1.0

Fourier transform based image registration

Foureg

Image registration using discrete Fourier transform.

Given two images, foureg calculates a similarity transformation that transforms one image into the other.

NOTE

THIS IS STILL WIP AND INTERFACES MAY CHANGE WITHOU NOTICE

Example

The example transforms an image with a user defined transformation and then rediscovers it using foureg.

import matplotlib.pyplot as plt
import numpy as np
from PIL import Image  # Not a dependency from this pa

from foureg import similarity, similarity_matrix, transform_img

# 1) Make up some transformation
transformation = similarity_matrix(1.2, 15, (40, 60))

# 2) Open the master image and transform it to generate the slave image
master = np.asarray(Image.open("./resources/examples/sample1.png"))
slave = transform_img(master, transformation)


# 3) Use foureg to recover the transformation
imreg_result = similarity(master, slave)
slave_transformed = transform_img(slave, imreg_result["transformation"])

4) Some plotting to verify everything is working
_, axs = plt.subplots(1, 5, figsize=(13, 8))
im_0 = axs[0].imshow(master)
plt.colorbar(im_0, ax=axs[0])
im_1 = axs[1].imshow(slave)
plt.colorbar(im_1, ax=axs[1])
im_2 = axs[2].imshow(slave_transformed)
plt.colorbar(im_2, ax=axs[2])
im_3 = axs[3].imshow(imreg_result["timg"])
plt.colorbar(im_3, ax=axs[3])
im_4 = axs[4].imshow(np.abs(imreg_result["timg"] - master))
plt.colorbar(im_4, ax=axs[4])

plt.show()

Features

• Image pre-processing options (frequency filtration, image extension).
• Under-the-hood options exposed (iterations, phase correlation filtration).
• Permissive open-source license (3-clause BSD).

Origin story

This is a fork of the imreg_dft borned of the desire to achieve the following goals:

• Ability to return the final transformation in matrix form as opposed to the angle, translation and scaling factor separately. The original code makes obtaining that matrix really hard because it does some unorthodox resizings when performing the image transformations.
• Better performance and ultimately a Pytorch powered GPU implementation
• A more focused codebase. The only goal here is to estimate similarity transformations between pairs of images.

Acknowledgements

The code was originally developed by Christoph Gohlke (University of California, Irvine, USA) and later on developed further by Matěj Týč (Brno University of Technology, CZ). This repo wouldn't exist without them.