03 – CPM Multi-Wavelength¶

This tutorial demonstrates polychromatic CPM reconstruction using a multi-wavelength dataset measured from a mouse brain sample. It shows how to handle spectral density weighting and run a multi-wavelength engine.

What you'll learn:

• How to set up a polychromatic (nlambda > 1) reconstruction
• How to use spectral density weighting
• How to interpret multi-wavelength reconstruction results

!!! note "Dataset" The dataset (Brain_smoothBeam_poly_bin8.hdf5, ~500 MB) is downloaded automatically into the example_data/ folder at the project root the first time you run the notebook. If the file already exists it is not re-downloaded.

!!! tip "Prerequisites" Complete 01 – CPM Simulation before this tutorial.

In [1]:

# %matplotlib notebook
import matplotlib
import matplotlib.pyplot as plt
import numpy as np
from pathlib import Path
import h5py

# %matplotlib notebook import matplotlib import matplotlib.pyplot as plt import numpy as np from pathlib import Path import h5py

In [2]:

# import the fracPy module
import PtyLab
from PtyLab import ExperimentalData
from PtyLab import Reconstruction
from PtyLab import Monitor
from PtyLab import Params
from PtyLab import Engines

# import the fracPy module import PtyLab from PtyLab import ExperimentalData from PtyLab import Reconstruction from PtyLab import Monitor from PtyLab import Params from PtyLab import Engines

In [3]:

from PtyLab.utils.visualisation import hsvplot, hsvmodeplot, absplot, absmodeplot

from PtyLab.utils.visualisation import hsvplot, hsvmodeplot, absplot, absmodeplot

In [ ]:

from PtyLab.io import getExampleDataFolder
import urllib.request

fileName = "Brain_smoothBeam_poly_bin8.hdf5"
filePath = getExampleDataFolder() / fileName

if not filePath.exists():
    print(f"Downloading {fileName}...")
    urllib.request.urlretrieve(
        "https://ndownloader.figshare.com/files/33950357",
        filePath,
    )
    print("Download complete.")
else:
    print(f"{fileName} found, skipping download.")

from PtyLab.io import getExampleDataFolder import urllib.request fileName = "Brain_smoothBeam_poly_bin8.hdf5" filePath = getExampleDataFolder() / fileName if not filePath.exists(): print(f"Downloading {fileName}...") urllib.request.urlretrieve( "https://ndownloader.figshare.com/files/33950357", filePath, ) print("Download complete.") else: print(f"{fileName} found, skipping download.")

Brain_smoothBeam_poly_bin8.hdf5 found, skipping download.

In [5]:

from PtyLab.utils.utils import posit
# Step 2: Initialize the ExperimentalData class

# Initialize the ExperimentalData using the filePath, and choose the operation Mode (default is 'CPM')
experimentalData = ExperimentalData(filePath, operationMode="CPM")
backgroundOffset = 10
experimentalData.entrancePupilDiameter = 0.2e-3

experimentalData.ptychogram = posit(experimentalData.ptychogram - backgroundOffset)
# show measured ptychogram in log scale. Need to close the window to continue
experimentalData.showPtychogram()

from PtyLab.utils.utils import posit # Step 2: Initialize the ExperimentalData class # Initialize the ExperimentalData using the filePath, and choose the operation Mode (default is 'CPM') experimentalData = ExperimentalData(filePath, operationMode="CPM") backgroundOffset = 10 experimentalData.entrancePupilDiameter = 0.2e-3 experimentalData.ptychogram = posit(experimentalData.ptychogram - backgroundOffset) # show measured ptychogram in log scale. Need to close the window to continue experimentalData.showPtychogram()

Found encoder with shape (890, 2)
Min max ptychogram: 0.0, 14378.9013671875
Min max ptychogram: 0.0, 4.157756328582764
0.0 4.1577563

VBox(children=(IntSlider(value=0, description='Frame', max=889), Output()))

In [12]:

# Set monitor properties
monitor = Monitor()
monitor.figureUpdateFrequency = 1
monitor.objectPlot = "angle"  # complex abs angle
monitor.verboseLevel = "low"  # high: plot two figures, low: plot only one figure
monitor.probeZoom = 1  # control probe plot FoV
monitor.objectZoom = 1  # control object plot FoV
monitor.objectPlotContrast = 8
monitor.probePlotContrast = 1

# Set monitor properties monitor = Monitor() monitor.figureUpdateFrequency = 1 monitor.objectPlot = "angle" # complex abs angle monitor.verboseLevel = "low" # high: plot two figures, low: plot only one figure monitor.probeZoom = 1 # control probe plot FoV monitor.objectZoom = 1 # control object plot FoV monitor.objectPlotContrast = 8 monitor.probePlotContrast = 1

/mnt/home/shantanu/projects/PtyLab.py/PtyLab/Monitor/Monitor.py:178: UserWarning: For faster update of the reconstruction plot, set `monitor.figureUpdateFrequency = 5` or higher.
  warnings.warn(

In [13]:

params = Params()
params.gpuSwitch = True  # This can be manually set to False if you want to run on CPU, but if GPU is available, it will be used by default
params.positionOrder = "random"  # 'sequential' or 'random'
params.propagatorType = (
    "Fraunhofer"  # Fraunhofer Fresnel ASP scaledASP polychromeASP scaledPolychromeASP
)
params.probePowerCorrectionSwitch = False
params.modulusEnforcedProbeSwitch = True
params.comStabilizationSwitch = True
params.orthogonalizationSwitch = True
params.orthogonalizationFrequency = 5
params.intensityConstraint = "standard"  # standard fluctuation exponential poission
params.couplingSwitch = True
params.couplingAleph = 0.5

params = Params() params.gpuSwitch = True # This can be manually set to False if you want to run on CPU, but if GPU is available, it will be used by default params.positionOrder = "random" # 'sequential' or 'random' params.propagatorType = ( "Fraunhofer" # Fraunhofer Fresnel ASP scaledASP polychromeASP scaledPolychromeASP ) params.probePowerCorrectionSwitch = False params.modulusEnforcedProbeSwitch = True params.comStabilizationSwitch = True params.orthogonalizationSwitch = True params.orthogonalizationFrequency = 5 params.intensityConstraint = "standard" # standard fluctuation exponential poission params.couplingSwitch = True params.couplingAleph = 0.5

INFO:GPU:cupy and CUDA available, switching to GPU

In [14]:

# now, all our experimental data is loaded into experimental_data and we don't have to worry about it anymore.
# now create an object to hold everything we're eventually interested in
reconstruction = Reconstruction(experimentalData, params)
reconstruction.No = 2**10
reconstruction.nlambda = len(reconstruction.spectralDensity)

# initialize probe and object and related params
reconstruction.initialProbe = "circ"
reconstruction.initialObject = "ones"
reconstruction.initializeObjectProbe()

# # customize initial probe quadratic phase
reconstruction.probe = reconstruction.probe * np.exp(
    1.0j
    * 2
    * np.pi
    / reconstruction.wavelength
    * (reconstruction.Xp**2 + reconstruction.Yp**2)
    / (1 * 2e-2)
)

hsvmodeplot(
    np.squeeze(reconstruction.probe), pixelSize=reconstruction.dxp, axisUnit="mm"
)

# now, all our experimental data is loaded into experimental_data and we don't have to worry about it anymore. # now create an object to hold everything we're eventually interested in reconstruction = Reconstruction(experimentalData, params) reconstruction.No = 2**10 reconstruction.nlambda = len(reconstruction.spectralDensity) # initialize probe and object and related params reconstruction.initialProbe = "circ" reconstruction.initialObject = "ones" reconstruction.initializeObjectProbe() # # customize initial probe quadratic phase reconstruction.probe = reconstruction.probe * np.exp( 1.0j * 2 * np.pi / reconstruction.wavelength * (reconstruction.Xp**2 + reconstruction.Yp**2) / (1 * 2e-2) ) hsvmodeplot( np.squeeze(reconstruction.probe), pixelSize=reconstruction.dxp, axisUnit="mm" )

INFO:Reconstruction:Copying attribute wavelength
INFO:Reconstruction:Copying attribute dxd
INFO:Reconstruction:Copying attribute theta
INFO:Reconstruction:Copying attribute spectralDensity
INFO:Reconstruction:Copying attribute entrancePupilDiameter
INFO:Reconstruction:Initial object set to ones
INFO:Reconstruction:Initial probe set to circ

In [15]:

# Run the reconstruction
## choose engine
# mPIE
multiPIE = Engines.multiPIE(reconstruction, experimentalData, params, monitor)
multiPIE.numIterations = 100
multiPIE.betaProbe = 0.25
multiPIE.betaObject = 0.25
multiPIE.reconstruct()

# Run the reconstruction ## choose engine # mPIE multiPIE = Engines.multiPIE(reconstruction, experimentalData, params, monitor) multiPIE.numIterations = 100 multiPIE.betaProbe = 0.25 multiPIE.betaObject = 0.25 multiPIE.reconstruct()

multiPIE: 100%|██████████| 100/100 [01:56<00:00,  1.17s/it]

INFO:multiPIE:switch to cpu

In [16]:

# now save the data
# reconstruction.saveResults('reconstruction.hdf5')

# now save the data # reconstruction.saveResults('reconstruction.hdf5')