Use next cell if too many logging messages…

import logging
LOG_FMT = '%(filename)s - l.%(lineno)d - %(levelname)s: %(message)s'
logging.basicConfig(format=LOG_FMT, level=logging.DEBUG)
logger = logging.getLogger()
logger.setLevel(logging.ERROR)

%matplotlib notebook
import matplotlib.pyplot as plt

from ipywidgets import interactive
from ipywidgets import IntSlider

def wimshow(slice_id):
    plt.imshow(blend[slice_id, :, :, :])
    plt.show()

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
Cell In[2], line 4
      1 get_ipython().run_line_magic('matplotlib', 'notebook')
      2 import matplotlib.pyplot as plt
----> 4 from ipywidgets import interactive
      5 from ipywidgets import IntSlider
      7 def wimshow(slice_id):

ModuleNotFoundError: No module named 'ipywidgets'

Blockmatching example with `vt-python`¶

In the following examples we try to register two angles of a multi-angle image (multiple observations of the same object at the same time but from different angles) using blockmatching.

import numpy as np
from timagetk.io.image import _image_from_url

/home/aurele/Soft/timagetk/src/timagetk/components/labelled_image.py:32: TqdmExperimentalWarning: Using `tqdm.autonotebook.tqdm` in notebook mode. Use `tqdm.tqdm` instead to force console mode (e.g. in jupyter console)
  from tqdm.autonotebook import tqdm

Defines location of shared data:¶

Input images (multi-angles)¶

flo_path = _image_from_url('https://zenodo.org/record/7149218/files/090223-p58-flo-tilt2.lsm',
                           hash_value='006bf3bacb667eb72d0fb9a6419edcc1', hash_method='md5')
ref_path = _image_from_url('https://zenodo.org/record/7149218/files/090223-p58-flo-top.lsm',
                           hash_value='5548917d3d1490200d0d56cbb31c0d35', hash_method='md5')

Here the reference image is top and the floating image is tilt2.

Please note that shared_data returns a string (str type in python) and do not read/load the actual file.

TODO: Add contour projection view of example images.

Output image & transformation¶

Using the command line calls to vt-python functions requires to defines file names and paths for the transformation matrix and the registered image.

trsf_rig_path = '/tmp/p58-t0-tilt2_on_top-rigid.trsf'
reg_rig_path = '/tmp/p58-t0-tilt2_on_top-rigid.tif'

Estimation of rigid transformation¶

It is now possible to estimate the rigid transformation registering image tilt2 onto top:

from vt import blockmatching as vt_blockmatching
vt_blockmatching(flo_path, ref_path, params="-trsf-type rigid -py-ll 2", out_file_path=trsf_rig_path)

Read the estimated rigid transformation¶

To access the estimated rigid transformation, you have to load it using the following commands:

from vt import vtTransformation as Trsf
rig_trsf = Trsf(trsf_rig_path)

You can now print a summary of the Trsf object:

print(rig_trsf)

Print the transformation matrix¶

To access the 4x4 matrix defining the transformation:

print(rig_trsf.copy_to_array())

Apply transformation to floating image¶

To obtain the registration of image tilt2 on the top reference, you now have to apply the previously estimated transformation rig_trsf to image tilt2. It is done as follow:

from vt import apply_trsf as vt_apply_trsf
vt_apply_trsf(flo_path, trsf_rig_path, ref=ref_path, params="-linear", out_file_path=reg_rig_path)

OPTIONAL - TimageTK - Visualization of the rigid registration¶

from timagetk.io import imread

flo_img = imread(flo_path)
ref_img = imread(ref_path)
reg_img = imread(reg_rig_path)

from timagetk.components.multi_channel import combine_channels

blend = combine_channels([ref_img, reg_img], ['green', 'red'])

interactive_plot = interactive(wimshow, slice_id=IntSlider(min=0, max=blend.get_shape('z')-1, step=1, value=0))
output = interactive_plot.children[-1]
interactive_plot

Estimation of affine transformation¶

trsf_aff_path = '/tmp/p58-t0-tilt2_on_top-affine.trsf'
reg_aff_path = '/tmp/p58-t0-tilt2_on_top-affine.tif'

from vt import blockmatching as vt_blockmatching
params = "-trsf-type affine -py-hl 5 -py-ll 1"
vt_blockmatching(flo_path, ref_path, params=params, out_file_path=trsf_aff_path)

from vt import apply_trsf as vt_apply_trsf
vt_apply_trsf(flo_path, trsf_aff_path, ref=ref_path, params="-linear", out_file_path=reg_aff_path)

reg_img = imread(reg_aff_path)

from timagetk.components.multi_channel import combine_channels

blend = combine_channels([ref_img, reg_img], ['green', 'red'])

interactive_plot = interactive(wimshow, slice_id=IntSlider(min=0, max=blend.get_shape('z')-1, step=1, value=0))
output = interactive_plot.children[-1]
interactive_plot

Estimation of deformable transformation¶

trsf_def_path = '/tmp/p58-t0-tilt2_on_top-deformable.trsf'
reg_def_path = '/tmp/p58-t0-tilt2_on_top-deformable.tif'

from vt import blockmatching as vt_blockmatching
params = "-trsf-type vectorfield -py-hl 5 -py-ll 1"
vt_blockmatching(flo_path, ref_path, trsf_init=trsf_aff_path, params=params, out_file_path=trsf_def_path)

from vt import apply_trsf as vt_apply_trsf
vt_apply_trsf(flo_path, trsf_def_path, ref=ref_path, params="-linear", out_file_path=reg_def_path)

reg_img = imread(reg_def_path)

from timagetk.components.multi_channel import combine_channels

blend = combine_channels([ref_img, reg_img], ['green', 'red'])

interactive_plot = interactive(wimshow, slice_id=IntSlider(min=0, max=blend.get_shape('z')-1, step=1, value=0))
output = interactive_plot.children[-1]
interactive_plot

Table of Contents

Blockmatching example with vt-python¶