Details and Options

• Local brightness adjustment is also known as flat fielding, and is used for removing image artifacts caused by nonuniform lighting or variations in sensor sensitivities.

• BrightnessEqualize works with arbitrary 2D and 3D images, adjusting the lightness channel in the LABColor space.
• A flatfield image is an image of a homogeneous signal like a plain well-lit white background. A darkfield is the same image obtained without lighting. The flat fielding of an image with an object in the same setting is given by .
• Possible settings for either flatfield or darkfield include:
• vala constant value val corrimagea correction image (rescaled to the image dimensions) {scope,model}fit the data into a given model
• The default flatfield consists of a 2^nd order polynomial fit. The default darkfield is assumed to be 0.
• Using {scope,model}, the flatfield or darkfield is estimated by fitting a function.
• The scope parameter specifies whether to fit the entire image data or the image projections along each axis. Possible settings include:
• "Global"fit the entire image to the model "Marginal"separately fit model to the projections along each axis
• The model can be one of the following:
• nan n-degree polynomial f,params,varsan arbitrary model f with parameters params and variables vars
• The following options are available:
• Masking Automaticthe regions to use for model estimation PerformanceGoalAutomaticaspects of performance to try to optimize
• Using Masking->Automatic, over- and underexposed pixels are not used for the adjustment.
• With partially transparent images, the alpha channel is multiplied with the mask.

Examples

Basic Examples  (1)

Equalize brightness of a plain texture:

Scope  (7)

Data  (3)

Equalize a grayscale image of a scanned document:

Adjust brightness in an unevenly illuminated color photo:

Adjust brightness of a 3D CT scan image:

Parameters  (4)

Remove uneven illumination in a microscopy image by estimating the brightness distribution:

Fitting a multivariate polynomial of first order on the entire image domain:

Using a second-order polynomial:

Fitting second-order univariate polynomials to the image marginals in each direction:

Using a flatfield image:

Remove a vignette effect by fitting the lightness channel to a given model:

Assuming a rotationally symmetric, radial-dependent vignette:

Assuming a Gaussian vignette:

Perform a subtractive correction on an astronomical image by estimating the dark field:

Options  (1)

Masking  (1)

By default, an automatic mask is used:

Equalize brightness based on the image background:

Applications  (7)

Achieve a constant luminosity of the Moon surface:

Equalize brightness of a microscopy image:

Identify the background:

Equalize brightness distribution assuming a homogeneous background:

Equalize color distribution by handling each channel separately:

Remove uneven illumination from a scanned document:

Create a mask to omit text in the subsequent equilibration:

Equalize brightness distribution fitting a sixth-order polynomial:

This improves the text recognition result:

Equalize the brightness distribution without a mask applying a horizontal fit:

This also improves the text recognition result:

Reduce the vignette effect of a photograph:

Compute a mask for the background that is expected to have even illumination:

Equalize image brightness:

Remove CCD artifacts from an image taken by a low-quality camera with many pixel defects.

Flat-field image of the camera:

Microscopy image of a tear drop:

Flat-field correction:

Correct uneven illumination in an MR scan:

Use only muscle tissue for brightness estimation:

Correct uneven brightness in a 3D magnetic resonance volume:

Note the reduced brightness toward the perimeter of the volume:

Select the muscle tissue to estimate the brightness distribution:

Estimate the brightness distribution by a polynomial in cylindrical coordinates of even orders up to 4:

Compare the initial volume and the result with equalized brightness: