histogram

histogram — find, manipulate and apply histograms and lookup tables

Stability Level

Stable, unless otherwise indicated

Functions

Includes

#include <vips/vips.h>

Description

Histograms and look-up tables are 1xn or nx1 images, where n is less than 256 or less than 65536, corresponding to 8- and 16-bit unsigned int images. They are tagged with a VipsInterpretation of VIPS_INTERPRETATION_HISTOGRAM and usually displayed by user-interfaces such as nip2 as plots rather than images.

These functions can be broadly grouped as things to find or build histograms (vips_hist_find(), vips_hist_find_indexed(), vips_hist_find_ndim(), vips_buildlut(), vips_identity()), operations that manipulate histograms in some way (vips_hist_cum(), vips_hist_norm()), operations to apply histograms (vips_maplut()), and a variety of utility operations.

A final group of operations build tone curves. These are useful in pre-press work for adjusting the appearance of images. They are designed for CIELAB images, but might be useful elsewhere.

Functions

vips_maplut ()

int
vips_maplut (VipsImage *in,
             VipsImage **out,
             VipsImage *lut,
             ...);

Optional arguments:

  • band : apply one-band lut to this band of in

Map an image through another image acting as a LUT (Look Up Table). The lut may have any type and the output image will be that type.

The input image will be cast to one of the unsigned integer types, that is, VIPS_FORMAT_UCHAR, VIPS_FORMAT_USHORT or VIPS_FORMAT_UINT.

If lut is too small for the input type (for example, if in is VIPS_FORMAT_UCHAR but lut only has 100 elements), the lut is padded out by copying the last element. Overflows are reported at the end of computation. If lut is too large, extra values are ignored.

If lut has one band and band is -1 (the default), then all bands of in pass through lut . If band is >= 0, then just that band of in passes through lut and other bands are just copied.

If lut has same number of bands as in , then each band is mapped separately. If in has one band, then lut may have many bands and the output will have the same number of bands as lut .

See also: vips_hist_find(), vips_identity().

[method]

Parameters

in

input image

 

out

output image.

[out]

lut

look-up table

 

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_percent ()

int
vips_percent (VipsImage *in,
              double percent,
              int *threshold,
              ...);

vips_percent() returns (through the threshold parameter) the threshold below which there are percent values of in . For example:

1
2
$ vips percent k2.jpg 90
214

Means that 90% of pixels in k2.jpg have a value less than 214.

The function works for uchar and ushort images only. It can be used to threshold the scaled result of a filtering operation.

See also: vips_hist_find(), vips_profile().

[method]

Parameters

in

input image

 

percent

threshold percentage

 

threshold

output threshold value.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_stdif ()

int
vips_stdif (VipsImage *in,
            VipsImage **out,
            int width,
            int height,
            ...);

Optional arguments:

  • a : weight of new mean

  • m0 : target mean

  • b : weight of new deviation

  • s0 : target deviation

vips_stdif() preforms statistical differencing according to the formula given in page 45 of the book "An Introduction to Digital Image Processing" by Wayne Niblack. This transformation emphasises the way in which a pel differs statistically from its neighbours. It is useful for enhancing low-contrast images with lots of detail, such as X-ray plates.

At point (i,j) the output is given by the equation:

1
2
vout(i,j) = @a * @m0 + (1 - @a) * meanv + 
      (vin(i,j) - meanv) * (@b * @s0) / (@s0 + @b * stdv)

Values a , m0 , b and s0 are entered, while meanv and stdv are the values calculated over a moving window of size width , height centred on pixel (i,j). m0 is the new mean, a is the weight given to it. s0 is the new standard deviation, b is the weight given to it.

Try:

1
vips stdif $VIPSHOME/pics/huysum.v fred.v 0.5 128 0.5 50 11 11

The operation works on one-band uchar images only, and writes a one-band uchar image as its result. The output image has the same size as the input.

See also: vips_hist_local().

[method]

Parameters

in

input image

 

out

output image.

[out]

width

width of region

 

height

height of region

 

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_cum ()

int
vips_hist_cum (VipsImage *in,
               VipsImage **out,
               ...);

Form cumulative histogram.

See also: vips_hist_norm().

[method]

Parameters

in

input image

 

out

output image.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_norm ()

int
vips_hist_norm (VipsImage *in,
                VipsImage **out,
                ...);

Normalise histogram. The maximum of each band becomes equal to the maximum index, so for example the max for a uchar image becomes 255. Normalise each band separately.

See also: vips_hist_cum().

[method]

Parameters

in

input image

 

out

output image.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_equal ()

int
vips_hist_equal (VipsImage *in,
                 VipsImage **out,
                 ...);

Optional arguments:

  • band : band to equalise

Histogram-equalise in . Equalise using band bandno , or if bandno is -1, equalise bands independently. The output format is always the same as the input format.

See also:

[method]

Parameters

in

input image

 

out

output image.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_plot ()

int
vips_hist_plot (VipsImage *in,
                VipsImage **out,
                ...);

Plot a 1 by any or any by 1 image file as a max by any or any by max image using these rules:

unsigned char max is always 256

other unsigned integer types output 0 - maxium value of in .

signed int types min moved to 0, max moved to max + min.

float types min moved to 0, max moved to any (square output)

[method]

Parameters

in

input image

 

out

output image.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_match ()

int
vips_hist_match (VipsImage *in,
                 VipsImage *ref,
                 VipsImage **out,
                 ...);

Adjust in to match ref . If in and ref are normalised cumulative histograms, out will be a LUT that adjusts the PDF of the image from which in was made to match the PDF of ref 's image.

See also: vips_maplut(), vips_hist_find(), vips_hist_norm(), vips_hist_cum().

[method]

Parameters

in

input histogram

 

ref

reference histogram

 

out

output histogram.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_local ()

int
vips_hist_local (VipsImage *in,
                 VipsImage **out,
                 int width,
                 int height,
                 ...);

Optional arguments:

  • max_slope : maximum brightening

Performs local histogram equalisation on in using a window of size width by height centered on the input pixel.

The output image is the same size as the input image. The edge pixels are created by mirroring the input image outwards.

If max_slope is greater than 0, it sets the maximum value for the slope of the cumulative histogram, that is, the maximum brightening that is performed. A value of 3 is often used. Local histogram equalization with contrast limiting is usually called CLAHE.

See also: vips_hist_equal().

[method]

Parameters

in

input image

 

out

output image.

[out]

width

width of region

 

height

height of region

 

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_ismonotonic ()

int
vips_hist_ismonotonic (VipsImage *in,
                       gboolean *out,
                       ...);

Test in for monotonicity. out is set non-zero if in is monotonic.

[method]

Parameters

in

lookup-table to test

 

out

set non-zero if in is monotonic.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_hist_entropy ()

int
vips_hist_entropy (VipsImage *in,
                   double *out,
                   ...);

Estimate image entropy from a histogram. Entropy is calculated as:

1
-sum( p * log2( p ) )

where p is histogram-value / sum-of-histogram-values.

[method]

Parameters

in

input histogram

 

out

image entropy.

[out]

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error


vips_case ()

int
vips_case (VipsImage *index,
           VipsImage **cases,
           VipsImage **out,
           int n,
           ...);

Use values in index to select pixels from cases .

index must have one band. cases can have up to 256 elements. Values in index greater than or equal to n use the final image in cases . The images in cases must have either one band or the same number of bands. The output image is the same size as index . Images in cases are expanded to the smallest common format and number of bands.

Combine this with vips_switch() to make something like a case statement or a multi-way vips_ifthenelse().

See also: vips_maplut(), vips_switch(), vips_ifthenelse().

[method]

Parameters

index

index image

 

cases

array of case images.

[array length=n]

out

output image.

[out]

n

number of case images

 

...

NULL-terminated list of optional named arguments

 

Returns

0 on success, -1 on error