stemtool.util package¶

Submodules¶

stemtool.util.fourier_reg module¶

stemtool.util.fourier_reg.dftregistration(buf1ft, buf2ft, usfac=1)[source]¶

Upsampled FFT registration between two images

Parameters:

buf1ft (ndarray) – Fourier transform of reference image, DC in (1,1) [DO NOT FFTSHIFT]
buf2ft (ndarray) – Fourier transform of image to register, DC in (1,1) [DO NOT FFTSHIFT]
usfac (int) – Upsampling factor (integer). Images will be registered to within 1/usfac of a pixel. For example usfac = 20 means the images will be registered within 1/20 of a pixel. (default = 1)

Returns:

row_shift (float) – Pixel shift in cartesian y direction
col_shift (float) – Pixel shift in cartesian x direction
error (float) – Translation invariant normalized RMS error between f and g
phase_diff (float) – Global phase difference between the two images (should be zero if images are non-negative).
registered_fft (ndarray) – Fourier transform of registered version of buf2ft, the global phase difference is compensated for.

Notes

Efficient subpixel image registration by crosscorrelation. This code gives the same precision as the FFT upsampled cross correlation in a small fraction of the computation time and with reduced memory requirements. It obtains an initial estimate of the crosscorrelation peak by an FFT and then refines the shift estimation by upsampling the DFT only in a small neighborhood of that estimate by means of a matrix-multiply DFT. With this procedure all the image points are used to compute the upsampled crosscorrelation.

References

[1]	Manuel Guizar-Sicairos, Samuel T. Thurman, and James R. Fienup, “Efficient subpixel image registration algorithms,” Opt. Lett. 33, 156-158 (2008).

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: * Redistributions of source code must retain the above copyright

notice, this list of conditions and the following disclaimer.

Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution
Neither the name of the University of Rochester nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Examples

If you have two images im1 and im2, run as:

>>> row_shift,col_shift,phase_diff,error,registered_fft = dftregistration(np.fft.fft2(im1),np.fft.fft2(im2),upsampling)

You can test by reversing the order

>>> row_shift_r,col_shift_r,phase_diff,error,registered_fft = dftregistration(np.fft.fft2(im2),np.fft.fft2(im1),upsampling)
>>> row_shift == -row_shift_r
>>> True
>>> col_shift == -col_shift_r
>>> True

stemtool.util.fourier_reg.dftups(input_image, usfac=1, nor=0, noc=0, roff=0, coff=0)[source]¶

Upsampled discrete Fourier transform

Parameters:	input_image (ndarray) – Input image usfac (int, optional) – Upsampling Factor. Default is 1 nor (int, optional) – Number of pixels in the output upsampled DFT, in units of upsampled pixels (default = size(in)) noc (int, optional) – Number of pixels in the output upsampled DFT, in units of upsampled pixels (default = size(in)) roff (int, optional) – Row offsets, allow to shift the output array to a region of interest on the DFT (default = 0) coff (int, optional) – Column offsets, allow to shift the output array to a region of interest on the DFT (default = 0)
Returns:	out_fft – Upsampled Fourier transform
Return type:	ndarray

Notes

Recieves DC in upper left corner, image center must be in [0,0] This code is intended to provide the same result as if the following operations were performed - Embed the array “input_image” in an array that is usfac times larger in each dimension. ifftshift to bring the center of the image to (1,1). - Take the FFT of the larger array - Extract an [nor, noc] region of the result. Starting with the [roff+1 coff+1] element. It achieves this result by computing the DFT in the output array without the need to zeropad. Much faster and memory efficient than the zero-padded FFT approach if [nor noc] are much smaller than [nr*usfac nc*usfac]

stemtool.util.fourier_reg.find_max_index(image)[source]¶

Find maxima in image

Parameters:	image (ndarray) – Input image
Returns:	ymax (int) – y-index position of maxima xmax (int) – x-index position of maxima

Notes

Finds the image maxima, and then locates the y and x indices corresponding to the maxima

Examples

>>> ym, xm = find_max_index(image)

stemtool.util.fourier_reg.first_max_index(image, order='C')[source]¶

First maxima in image

Parameters:

image (ndarray) – Input image
order ({'C','F', 'A', 'K'},) – optional The elements of a are read using this index order. ‘C’ means to index the elements in row-major, C-style order, with the last axis index changing fastest, back to the first axis index changing slowest. ‘F’ means to index the elements in column-major, Fortran-style order, with the first index changing fastest, and the last index changing slowest. Note that the ‘C’ and ‘F’ options take no account of the memory layout of the underlying array, and only refer to the order of axis indexing. ‘A’ means to read the elements in Fortran-like index order if a is Fortran contiguous in memory, C-like order otherwise. ‘K’ means to read the elements in the order they occur in memory, except for reversing the data when strides are negative. By default, ‘C’ index order is used.

Returns:

ymax (int) – y-index position of maxima
xmax (int) – x-index position of maxima

Notes

Finds the first image maxima if there are multiple points with the same maximum value, and then locates the y and x indices corresponding to the maxima

Examples

>>> ym, xm = first_max_index(image)

stemtool.util.fourier_reg.fourier_pad(imFT, outsize)[source]¶

Pad Fourier images

Parameters:	imFT (ndarray) – Input complex array with DC in [1,1] outsize (ndarray with (2,1) shape) – ny, nx of output size
Returns:	imout – Output complex image with DC in [1,1]
Return type:	ndarray

Notes

Pads or crops the Fourier transform to the desired ouput size. Taking care that the zero frequency is put in the correct place for the output for subsequent FT or IFT. Can be used for Fourier transform based interpolation, i.e. dirichlet kernel interpolation.

stemtool.util.gauss_utils module¶

stemtool.util.gauss_utils.fit_gaussian1D_mask(signal, position, mask_width, center_type='COM')[source]¶

Fit a 2D gaussian to a masked image based on the location of the mask, size of the mask and the type of the mask

Parameters:	signal (ndarray) – The 1D signal that will be fitted with the Gaussian position (float) – x center of the mask mask_width (float) – The size of the mask. center_type (str) – Center location for the first pass of the Gaussian. Default is COM, while the other options are minima or maxima.
Returns:	popt – Refined X position, Standard deviation, Amplitude
Return type:	tuple

Notes

This code uses the scipy.optimize.curve_fit module to fit a 1D Gaussian peak to masked data. mask_x refers to the initial starting position. Also, this can take in minima as a string for initializing Gaussian peaks.

See also

fit_gaussian2D_mask initialize_gauss1D gaussian_1D_function

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.gauss_utils.fit_gaussian2D_mask(image_data, mask_x, mask_y, mask_radius, mask_type='circular', center_type='COM')[source]¶

Fit a 2D gaussian to a masked image based on the location of the mask, size of the mask and the type of the mask

Parameters:	image_data (ndarray) – The image that will be fitted with the Gaussian mask_x (float) – x center of the mask mask_y (float) – y center of the mask mask_radius (float) – The size of the mask. For a circulat mask this refers to the mask radius, while for a square mask this refers to half the side of the square mask_type (str) – Default is circular, while the other option is square center_type (str) – Center location for the first pass of the Gaussian. Default is COM, while the other options are minima or maxima.
Returns:	popt – Refined X position, Refined Y Position, Rotation angle of 2D Gaussian, Standard deviation(s), Amplitude
Return type:	tuple

Notes

This code uses the scipy.optimize.curve_fit module to fit a 2D Gaussian peak to masked data. mask_x and mask_y refer to the initial starting positions. Also, this can take in minima as a string for initializing Gaussian peaks, which allows for atom column mapping in inverted contrast images too.

See also

gaussian_2D_function initialize_gauss2D

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.gauss_utils.gauss2D(im_size, x0, y0, theta, sigma_x, sigma_y, amplitude)[source]¶

Return a 2D Gaussian function centered at x0, y0

Parameters:	im_size (tuple) – Size of the image where a Gaussian peak is to be generated x0 (float) – x center of Gaussian peak y0 (float) – y center of Gaussian peak theta (float) – Rotation of the 2D gaussian peak in radians sigma_x (float) – Standard deviation of the 2D Gaussian along x sigma_y (float) – Standard deviation of the 2D Gaussian along y amplitude (float) – Peak intensity
Returns:	gauss2D – 2D Gaussian peak of shape im_size
Return type:	ndarray

Notes

This returns a 2D ndarray with the the size as im_size, with a 2D gaussian function centered at (x0,y0) defined by the input parameters.

See also

gauss2D_function

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.gauss_utils.gaussian_1D_function(x, x0, sigma_x, amplitude)[source]¶

The underlying 1D Gaussian function

Parameters:	x (ndarray) – x positions x0 (float) – x center of Gaussian peak sigma_x (float) – Standard deviation of the 2D Gaussian along x amplitude (float) – Peak intensity
Returns:	gaussvals – A Gausian peak centered at x0 based on the parameters given
Return type:	ndarray

Notes

The Gaussian 1D function is calculated at every x position for a list of position values based on the parameters given.

stemtool.util.gauss_utils.gaussian_2D_function(xy, x0, y0, theta, sigma_x, sigma_y, amplitude)[source]¶

The underlying 2D Gaussian function

Parameters:	xy (tuple) – x and y positions x0 (float) – x center of Gaussian peak y0 (float) – y center of Gaussian peak theta (float) – Rotation of the 2D gaussian peak in radians sigma_x (float) – Standard deviation of the 2D Gaussian along x sigma_y (float) – Standard deviation of the 2D Gaussian along y amplitude (float) – Peak intensity
Returns:	gaussvals – A Gausian peak centered at x0, y0 based on the parameters given
Return type:	ndarray

Notes

The Gaussian 2D function is calculated at every x and y position for a list of position values based on the parameters given.

See also

gauss2D

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.gauss_utils.initialize_gauss1D(xx, yy, center_type='COM')[source]¶

Generate an approximate Gaussian based on signal

Parameters:	xx (ndarray) – X positions yy (ndarray) – Signal value at the positions center_type (str) – Default is COM which uses the center of mass of the given positions to generate the starting gaussian center. The other option is maxima which takes the maximum intensity value as the starting point.
Returns:	gauss_ini – X_center, X_std, Amplitude
Return type:	tuple

Notes

For a given list of x positions and corresponding signal values, this code returns a first pass approximation of a 1D Gaussian function. The center of the function can either be the center of mass or the intensity maxima, and is user defined.

stemtool.util.gauss_utils.initialize_gauss2D(xx, yy, zz, center_type='COM')[source]¶

Generate an approximate Gaussian based on image

Parameters:	xx (ndarray) – X positions yy (ndarray) – Y Positions zz (ndarray) – Image value at the positions center_type (str) – Default is COM which uses the center of mass of the given positions to generate the starting gaussian center. The other option is maxima which takes the maximum intensity value as the starting point.
Returns:	gauss_ini – X_center, Y_center, Angle, X_std, Y_std, Amplitude
Return type:	tuple

Notes

For a given list of x positions, y positions and corresponding intensity values, this code returns a first pass approximation of a Gaussian function. The center of the gaussian 2D function can either be the center of mass or the intensity maxima, and is user defined. The angle is always given as 0.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils module¶

stemtool.util.image_utils.array_rms(arr)[source]¶

stemtool.util.image_utils.circle_function(xy, x0, y0, radius)[source]¶

stemtool.util.image_utils.cp_image_sat(comp_image)[source]¶

stemtool.util.image_utils.cp_image_val(comp_image)[source]¶

stemtool.util.image_utils.cross_corr(image_1, image_2, hybridizer=0, normal=True)[source]¶

Normalized Correlation, allowing for hybridization with cross correlation being the default output if no hybridization parameter is given

image_1: ndarray

First image

image_2: ndarray

Second image

hybridizer: float

Hybridization parameter between 0 and 1 0 is pure cross correlation 1 is pure phase correlation

corr_hybrid: ndarray

Complex valued correlation output

The cross-correlation theorem can be stated as: .. math:
G_c = G_1      imes G_2^*
where \(G_c\) is the Fourier transform of the cross correlated matrix and \(G_1\) and \(G_2\) are the Fourier transforms of \(g_1\) and \(g_2\) respectively, which are the original matrices. This is pure cross-correlation. Phase correlation can be expressed as: .. math:
G_c =

rac{G_1 imes G_2^*}{mid G_1 imes G_2^* mid}

Thus, we can now define a hybrid cross-correlation where .. math:
G_c =

rac{G_1 imes G_2^*}{mid G_1 imes G_2^* mid ^n}

If n is 0, we have cross correlation, and if n is 1 we have phase correlation.

1]_, Pekin, T.C., Gammer, C., Ciston, J., Minor, A.M. and Ophus, C.,

2017. Optimizing disk registration algorithms for nanobeam electron diffraction strain mapping. Ultramicroscopy, 176, pp.170-176.

sparse_division

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.cross_corr_unpadded(image_1, image_2, normal=True)[source]¶

stemtool.util.image_utils.euclidean_dist(binary_image)[source]¶

Get Euclidean distance of every non-zero pixel from zero valued pixels

Parameters:	binary_image (ndarray) – The image
Returns:	dist_map – The Euclidean distance map
Return type:	ndarray

stemtool.util.image_utils.fit_circle(image_data, med_factor=50)[source]¶

stemtool.util.image_utils.get_mean_std(xlist, ylist, resolution=25, style='median')[source]¶

Get mean and standard deviation of a list of x and y values

Parameters:

xlist (ndarray) – (n,1) shape of x values
ylist (ndarray) – (n,1) shape of x values
resolution (float) – Optional value, default if 50 How finely sampled the final list is

Returns:

stdvals (ndarray) – First column is x values sampled at resolution Second Column is mean or median of corresponding y values Third column is the standard deviation
Authors:
Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.hanned_image(image)[source]¶

2D hanning filter for images

Parameters:	image (ndarray) – Original Image on which the Hanning filter is to be applied
Returns:	hanned_image – Image with the hanning filter applied
Return type:	ndarray

Notes

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.hsv_overlay(data, image, color, climit=None, bit_depth=8)[source]¶

stemtool.util.image_utils.image_logarizer(image_orig, bit_depth=64)[source]¶

Normalized log of image

Parameters:	image_orig (ndarray) – Numpy array of real valued image bit_depth (int) – Bit depth of output image Default is 32
Returns:	image_log – Normalized log
Return type:	ndarray

Notes

Normalize the image, and scale it 2^0 to 2^bit_depth. Take log2 of the scaled image.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.image_normalizer(image_orig)[source]¶

Normalizing Image

Parameters:	image_orig (ndarray) – ‘image_orig’ is the original input image to be normalized
Returns:	image_norm – Normalized Image
Return type:	ndarray

Notes

We normalize a real valued image here so that it’s values lie between o and 1. This is done by first subtracting the minimum value of the image from the image itself, and then subsequently dividing the image by the maximum value of the subtracted image.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.is_odd(num)[source]¶

stemtool.util.image_utils.make_circle(size_circ, center_x, center_y, radius)[source]¶

Make a circle

Parameters:

size_circ (ndarray) – 2 element array giving the size of the output matrix
center_x (float) – x position of circle center
center_y (float) – y position of circle center
radius (float) – radius of the circle

Returns:

circle (ndarray) – p X q sized array where the it is 1 inside the circle and 0 outside
Authors:
Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.move_by_phase(image_to_move, x_pixels, y_pixels)[source]¶

Move Images with sub-pixel precision

Parameters:	image_to_move (ndarray) – Original Image to be moved x_pixels (float) – Pixels to shift in X direction y_pixels (float) – Pixels to Shift in Y direction
Returns:	moved_image – Moved Image
Return type:	ndarray

Notes

The underlying idea is that a shift in the real space is phase shift in Fourier space. So we take the original image, and take it’s Fourier transform. Also, we calculate how much the image shifts result in the phase change, by calculating the Fourier pixel dimensions. We then multiply the FFT of the image with the phase shift value and then take the inverse FFT.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.phase_color(phase_image)[source]¶

stemtool.util.image_utils.reduce_precision_xy(xy, reducer)[source]¶

Reduce the precison of an xy array along both the x and y axes.

Parameters:	xy (ndarray) – The first column is the x dimension, while the second column is the y dimension reducer (tuple) – The precison reducers along the individual dimensions
Returns:	xyprec – Reduced precision array
Return type:	ndarray

stemtool.util.image_utils.remove_dead_pixels(image_orig, iter_count=1, level=10000)[source]¶

Remove dead pixels

Parameters:	image_orig (ndarray) – Numpy array of real valued image iter_count (int) – Number of iterations to run the process. Default is 1 level (int,float) – Ratio of minima pixels to total pixels. Default is 10,000
Returns:	image_orig – Image with dead pixels converted
Return type:	ndarray

Notes

Subtract the minima from the image, and if the number of pixels with minima values is less than the 1/level of the total pixels, then those are decided to be dead pixels. Iterate if necessary

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.sane_colorbar(mappable)[source]¶

stemtool.util.image_utils.sobel_circle(image)[source]¶

stemtool.util.image_utils.sparse_division(sparse_numer, sparse_denom, bit_depth=32)[source]¶

Divide two sparse matrices element wise to prevent zeros

Parameters:	spase_numer (ndarray) – Numpy array of real valued numerator sparse_denom (ndarray) – Numpy array of real valued denominator bit_depth (int) – Bit depth of output image Default is 32
Returns:	divided_matrix – Quotient matrix
Return type:	ndarray

Notes

Decide on a bit depth below which the values in the denominator are just noise, as they are below the bit depth. Do the same for the numerator. Turn those values to 1 in the denominator and 0 in the numerator. Then in the quotient matrix, turn the denominator values below the threshold to 0 too.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.pnccd module¶

stemtool.util.image_utils.array_rms(arr)[source]

stemtool.util.image_utils.circle_function(xy, x0, y0, radius)[source]

stemtool.util.image_utils.cp_image_sat(comp_image)[source]

stemtool.util.image_utils.cp_image_val(comp_image)[source]

stemtool.util.image_utils.cross_corr(image_1, image_2, hybridizer=0, normal=True)[source]

Normalized Correlation, allowing for hybridization with cross correlation being the default output if no hybridization parameter is given

image_1: ndarray

First image

image_2: ndarray

Second image

hybridizer: float

Hybridization parameter between 0 and 1 0 is pure cross correlation 1 is pure phase correlation

corr_hybrid: ndarray

Complex valued correlation output

The cross-correlation theorem can be stated as: .. math:
G_c = G_1      imes G_2^*
where \(G_c\) is the Fourier transform of the cross correlated matrix and \(G_1\) and \(G_2\) are the Fourier transforms of \(g_1\) and \(g_2\) respectively, which are the original matrices. This is pure cross-correlation. Phase correlation can be expressed as: .. math:
G_c =

rac{G_1 imes G_2^*}{mid G_1 imes G_2^* mid}

Thus, we can now define a hybrid cross-correlation where .. math:
G_c =

rac{G_1 imes G_2^*}{mid G_1 imes G_2^* mid ^n}

If n is 0, we have cross correlation, and if n is 1 we have phase correlation.

1]_, Pekin, T.C., Gammer, C., Ciston, J., Minor, A.M. and Ophus, C.,

2017. Optimizing disk registration algorithms for nanobeam electron diffraction strain mapping. Ultramicroscopy, 176, pp.170-176.

sparse_division

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.cross_corr_unpadded(image_1, image_2, normal=True)[source]

stemtool.util.image_utils.euclidean_dist(binary_image)[source]

Get Euclidean distance of every non-zero pixel from zero valued pixels

Parameters:	binary_image (ndarray) – The image
Returns:	dist_map – The Euclidean distance map
Return type:	ndarray

stemtool.util.image_utils.fit_circle(image_data, med_factor=50)[source]

stemtool.util.image_utils.get_mean_std(xlist, ylist, resolution=25, style='median')[source]

Get mean and standard deviation of a list of x and y values

Parameters:

xlist (ndarray) – (n,1) shape of x values
ylist (ndarray) – (n,1) shape of x values
resolution (float) – Optional value, default if 50 How finely sampled the final list is

Returns:

stdvals (ndarray) – First column is x values sampled at resolution Second Column is mean or median of corresponding y values Third column is the standard deviation
Authors:
Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.hanned_image(image)[source]

2D hanning filter for images

Parameters:	image (ndarray) – Original Image on which the Hanning filter is to be applied
Returns:	hanned_image – Image with the hanning filter applied
Return type:	ndarray

Notes

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.hsv_overlay(data, image, color, climit=None, bit_depth=8)[source]

stemtool.util.image_utils.image_logarizer(image_orig, bit_depth=64)[source]

Normalized log of image

Parameters:	image_orig (ndarray) – Numpy array of real valued image bit_depth (int) – Bit depth of output image Default is 32
Returns:	image_log – Normalized log
Return type:	ndarray

Notes

Normalize the image, and scale it 2^0 to 2^bit_depth. Take log2 of the scaled image.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.image_normalizer(image_orig)[source]

Normalizing Image

Parameters:	image_orig (ndarray) – ‘image_orig’ is the original input image to be normalized
Returns:	image_norm – Normalized Image
Return type:	ndarray

Notes

We normalize a real valued image here so that it’s values lie between o and 1. This is done by first subtracting the minimum value of the image from the image itself, and then subsequently dividing the image by the maximum value of the subtracted image.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.is_odd(num)[source]

stemtool.util.image_utils.make_circle(size_circ, center_x, center_y, radius)[source]

Make a circle

Parameters:

size_circ (ndarray) – 2 element array giving the size of the output matrix
center_x (float) – x position of circle center
center_y (float) – y position of circle center
radius (float) – radius of the circle

Returns:

circle (ndarray) – p X q sized array where the it is 1 inside the circle and 0 outside
Authors:
Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.move_by_phase(image_to_move, x_pixels, y_pixels)[source]

Move Images with sub-pixel precision

Parameters:	image_to_move (ndarray) – Original Image to be moved x_pixels (float) – Pixels to shift in X direction y_pixels (float) – Pixels to Shift in Y direction
Returns:	moved_image – Moved Image
Return type:	ndarray

Notes

The underlying idea is that a shift in the real space is phase shift in Fourier space. So we take the original image, and take it’s Fourier transform. Also, we calculate how much the image shifts result in the phase change, by calculating the Fourier pixel dimensions. We then multiply the FFT of the image with the phase shift value and then take the inverse FFT.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.phase_color(phase_image)[source]

stemtool.util.image_utils.reduce_precision_xy(xy, reducer)[source]

Reduce the precison of an xy array along both the x and y axes.

Parameters:	xy (ndarray) – The first column is the x dimension, while the second column is the y dimension reducer (tuple) – The precison reducers along the individual dimensions
Returns:	xyprec – Reduced precision array
Return type:	ndarray

stemtool.util.image_utils.remove_dead_pixels(image_orig, iter_count=1, level=10000)[source]

Remove dead pixels

Parameters:	image_orig (ndarray) – Numpy array of real valued image iter_count (int) – Number of iterations to run the process. Default is 1 level (int,float) – Ratio of minima pixels to total pixels. Default is 10,000
Returns:	image_orig – Image with dead pixels converted
Return type:	ndarray

Notes

Subtract the minima from the image, and if the number of pixels with minima values is less than the 1/level of the total pixels, then those are decided to be dead pixels. Iterate if necessary

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>

stemtool.util.image_utils.sane_colorbar(mappable)[source]

stemtool.util.image_utils.sobel_circle(image)[source]

stemtool.util.image_utils.sparse_division(sparse_numer, sparse_denom, bit_depth=32)[source]

Divide two sparse matrices element wise to prevent zeros

Parameters:	spase_numer (ndarray) – Numpy array of real valued numerator sparse_denom (ndarray) – Numpy array of real valued denominator bit_depth (int) – Bit depth of output image Default is 32
Returns:	divided_matrix – Quotient matrix
Return type:	ndarray

Notes

Decide on a bit depth below which the values in the denominator are just noise, as they are below the bit depth. Do the same for the numerator. Turn those values to 1 in the denominator and 0 in the numerator. Then in the quotient matrix, turn the denominator values below the threshold to 0 too.

Authors:

Debangshu Mukherjee <mukherjeed@ornl.gov>