Processing Images#

This section describes the image processing features available in DataLab.

Geometry#

Flip and rotation#

Create a new image by flipping or rotating the data of the selected image. The image may be flipped horizontally, vertically, or diagonally (transposition). It may be rotated by 90°, 270° or any user-defined value.

Operation	Description
Flip horizontally	Mirror image along the vertical axis
Flip diagonally	Transpose the image (swap X/Y axes)
Flip vertically	Mirror image along the horizontal axis
Rotate 90° right	Rotate image 90° clockwise
Rotate 90° left	Rotate image 90° counter-clockwise
Rotate by…	Rotate image by any angle (user-defined)

Distribute images along a grid#

Feature	Description
Distribute on a grid	Distribute selected images on a regular grid
Reset image positions	Reset the positions of the selected images to first image (x0, y0) coordinates

Axis transformation#

Set uniform coordinates#

Create a new image with uniform coordinates (i.e., with a constant pixel size in both X and Y directions).

Polynomial calibration#

Create a new image which is a polynomial calibration of each selected image with respect to Z axis.

Note

Linear calibration was removed from the UI as it is a special case of polynomial calibration with degree 1.

Level adjustment#

Normalize#

Create a new image which is the normalized version of each selected image by maximum, amplitude, sum, energy or RMS:

Normalization	Equation
Maximum	\(z_{1} = \dfrac{z_{0}}{z_{\max}}\)
Amplitude	\(z_{1} = \dfrac{z_{0}}{z_{\max}-z_{\min}}\)
Area	\(z_{1} = \dfrac{z_{0}}{\sum_{i=0}^{N-1}{z_{i}}}\)
Energy	\(z_{1}= \dfrac{z_{0}}{\sqrt{\sum_{n=0}^{N}\left\|z_{0}[n]\right\|^2}}\)
RMS	\(z_{1}= \dfrac{z_{0}}{\sqrt{\dfrac{1}{N}\sum_{n=0}^{N}\left\|z_{0}[n]\right\|^2}}\)

Clipping#

Apply the clipping to each selected image.

Offset correction#

Create a new image which is the result of offset correction on each selected image. This operation is performed by subtracting the image background value which is estimated by the mean value of a user-defined rectangular area.

Noise addition#

Generate new images by adding the same noise to each selected image. The available noise types are:

Noise	Description
Gaussian	Normal distribution
Uniform	Uniform distribution
Poisson	Poisson distribution

Noise reduction#

Create a new image which is the result of noise reduction on each selected image.

The following filters are available:

Filter	Formula/implementation
Gaussian filter	scipy.ndimage.gaussian_filter
Moving average	scipy.ndimage.uniform_filter
Moving median	scipy.ndimage.median_filter
Wiener filter	scipy.signal.wiener

Fourier analysis#

Zero padding#

Create a new image which is the result of zero padding on each selected image.

The following parameters are available:

Parameter	Description
Strategy	Zero padding strategy (see below)
Rows	Number of rows to add (if strategy is ‘custom’)
Columns	Number of columns to add (if strategy is ‘custom’)
Position	Position of the added zeros: ‘bottom-right’, ‘centered’

Zero padding strategy refers to the method used to add zeros to the image, and it can be one of the following:

Strategy	Description
next_pow2	Next power of 2 (e.g. 512, 1024, …)
multiple_of_64	Next multiple of 64 (e.g. 512, 576, …)
custom	Custom size (user-defined)

FFT related functions#

Create a new image which is the result of a Fourier analysis on each selected image.

The following functions are available:

Function	Description	Formula/implementation
FFT	Fast Fourier Transform	numpy.fft.fft2
Inverse FFT	Inverse Fast Fourier Transform	numpy.fft.ifft2
Magnitude spectrum	Optional: output in decibels (dB)	\(z_{1} = \left\|\FFT\left(z_{0}\right)\right\|\) or \(z_{1} = 20 \log_{10} \left(\left\|\FFT\left(z_{0}\right)\right\|\right)\) (dB)
Phase spectrum	Phase of the FFT expressed in degrees, using numpy.angle	\(z_{1} = \angle\left(\FFT\left(z_{0}\right)\right)\)
Power spectral density	Optional: output in decibels (dB)	\(z_{1} = \left\|\FFT\left(z_{0}\right)\right\|^2\) or \(z_{1} = 10 \log_{10} \left(\left\|\FFT\left(z_{0}\right)\right\|^2\right)\) (dB)

Note

FFT and inverse FFT are performed using frequency shifting if the option is enabled in DataLab settings (see Settings).

Frequency-domain filters#

The following frequency-domain filters are available:

Method	Description
Butterworth	Butterworth filter, based on skimage.filters.butterworth
Gaussian filter	Gaussian filter

Thresholding#

Create a new image which is the result of thresholding on each selected image, eventually based on user-defined parameters (“Parametric thresholding”).

The following parameters are available when selecting “Parametric thresholding”:

Parameter	Description
Threshold method	The thresholding method to use (see table below)
Bins	Number of bins for histogram calculation
Value	Threshold value
Operation	Operation to apply (> or <)

The following thresholding methods are available:

Method	Implementation
Manual	Manual thresholding (user-defined parameters)
ISODATA	skimage.filters.threshold_isodata
Li	skimage.filters.threshold_li
Mean	skimage.filters.threshold_mean
Minimum	skimage.filters.threshold_minimum
Otsu	skimage.filters.threshold_otsu
Triangle	skimage.filters.threshold_triangle
Yen	skimage.filters.threshold_yen

Note

The “All thresholding methods” option allows to perform all thresholding methods on the same image. Combined with the “distribute on a grid” option, this allows to compare the different thresholding methods on the same image.

Exposure#

Create a new image which is the result of exposure correction on each selected image.

The following functions are available:

Function	Implementation	Comments
Gamma correction	skimage.exposure.adjust_gamma
Logarithmic correction	skimage.exposure.adjust_log
Sigmoid correction	skimage.exposure.adjust_sigmoid
Histogram equalization	skimage.exposure.equalize_hist
Adaptive histogram equalization	skimage.exposure.equalize_adapthist	Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm
Intensity rescaling	skimage.exposure.rescale_intensity	Stretch or shrink image intensity levels

Restoration#

Create a new image which is the result of restoration on each selected image.

The following functions are available:

Function	Implementation	Comments
Total variation denoising	skimage.restoration.denoise_tv_chambolle
Bilateral filter denoising	skimage.restoration.denoise_bilateral
Wavelet denoising	skimage.restoration.denoise_wavelet
White Top-Hat denoising	skimage.morphology.white_tophat	Denoise image by subtracting its white top hat transform

Note

The “All denoising methods” option allows to perform all denoising methods on the same image. Combined with the “distribute on a grid” option, this allows to compare the different denoising methods on the same image.

Morphology#

Create a new image which is the result of morphological operations on each selected image, using a disk footprint.

The following functions are available:

Function	Implementation
White Top-Hat (disk)	skimage.morphology.white_tophat
Black Top-Hat (disk)	skimage.morphology.black_tophat
Erosion (disk)	skimage.morphology.erosion
Dilation (disk)	skimage.morphology.dilation
Opening (disk)	skimage.morphology.opening
Closing (disk)	skimage.morphology.closing

Note

The “All morphological operations” option allows to perform all morphological operations on the same image. Combined with the “distribute on a grid” option, this allows to compare the different morphological operations on the same image.

Edges#

Create a new image which is the result of edge filtering on each selected image.

The following functions are available:

Function	Implementation
Canny filter	skimage.feature.canny
Farid filter	skimage.filters.farid
Farid filter (horizontal)	skimage.filters.farid_h
Farid filter (vertical)	skimage.filters.farid_v
Laplace filter	skimage.filters.laplace
Prewitt filter	skimage.filters.prewitt
Prewitt filter (horizontal)	skimage.filters.prewitt_h
Prewitt filter (vertical)	skimage.filters.prewitt_v
Roberts filter	skimage.filters.roberts
Scharr filter	skimage.filters.scharr
Scharr filter (horizontal)	skimage.filters.scharr_h
Scharr filter (vertical)	skimage.filters.scharr_v
Sobel filter	skimage.filters.sobel
Sobel filter (horizontal)	skimage.filters.sobel_h
Sobel filter (vertical)	skimage.filters.sobel_v

Note

The “All edges filters” option allows to perform all edge filtering algorithms on the same image. Combined with the “distribute on a grid” option, this allows to compare the different edge filters on the same image.

Erase area#

Erase an area in the image as defined by a region of interest (ROI).

Note

The region to erase is defined by the user through a dialog box. It can consist of a single ROI or multiple ROIs with various shapes (rectangular, circular, etc.). Note that the ROI defined here is not bound to any object; it is used solely to specify the area to erase in the image. In particular, it is independent of the image’s ROI, if any (the latter is shown in the dialog box as a masked area).

Resize#

Create a new image which is a resized version of each selected image.

Pixel binning#

Combine clusters of adjacent pixels, throughout the image, into single pixels. The result can be the sum, average, median, minimum, or maximum value of the cluster.

Resampling#

Generate new images by resampling each selected image. The following parameters are available:

Parameter	Description
math:x_{min}	Minimum x-coordinate of the output image
\(x_{max}\)	Maximum x-coordinate of the output image
\(y_{min}\)	Minimum y-coordinate of the output image
\(y_{max}\)	Maximum y-coordinate of the output image
Mode	Image size definition mode: ‘Pixel size’ or ‘Output shape’. The ‘Pixel size’ mode allows to define the pixel size of the new image, while the ‘Output shape’ mode allows to define the number of pixels of the new image.
ΔX	Pixel size in x-direction (if ‘Pixel size’ mode is selected)
ΔY	Pixel size in y-direction (if ‘Pixel size’ mode is selected)
Width	Output image width in pixels (if ‘Output shape’ mode is selected)
Height	Output image height in pixels (if ‘Output shape’ mode is selected)
Interpolation method	Interpolation method to use: ‘nearest’, ‘linear’, ‘cubic’
Fill value	Value to use for points outside the input image domain (if None, function uses NaN for extrapolation)