Histograms

Functions
array	histogram (const array &data, unsigned nbins)
	Histogram of all values in data.
array	histogram (const array &data, unsigned nbins, float min, float max)
	Histogram of all values in data.
array	histequal (const array &data, const array &histogram)
	Data normalization via histogram equalization.

---

Function Documentation

array af::histogram	(	const array &	data,
		unsigned	nbins
	)

Histogram of all values in data.

       // create image histogram using 256 bins
       array ihist = histogram(img_in, 256);

Parameters:

[in]	data	image treated as vector
[in]	nbins	Number of output bins to populate

Returns:

non-normalized histogram of data (1D)

Examples:

examples/image_processing/image_demo.cpp, and examples/machine_learning/kmeans.cpp.

array af::histogram	(	const array &	data,
		unsigned	nbins,
		float	min,
		float	max
	)

Histogram of all values in data.

       // create image histogram using 100 bins,
       //  with bins scaled from 1 to 10, and
       //   bin[0]  gets -inf to 1
       //   bin[99] gets   10 to inf
       array ihist = histogram(img_in, 100, 0, 10);

Parameters:

[in]	data	image treated as vector
[in]	nbins	Number of bins to populate between min and max
[in]	min	minimum bin value (accumulates -inf to min)
[in]	max	maximum bin value (accumulates max to inf)

Returns:

non-normalized histogram of data (1D)

array af::histequal	(	const array &	data,
		const array &	histogram
	)

Data normalization via histogram equalization.

       // img_in and img_out are single channel images
       // image histogram equalization
       array ihist = histogram(img_in, 256);
       array img_out = histequal(img_in, ihist);

Parameters:

[in]	data	non-normalized input (!! assumes values [0-255] !!)
[in]	histogram	target histogram to approximate in output (based on # of bins)

Returns:

data with histogram approximately equal to histogram

Note:

image must be grayscale (single channel)

Examples:

examples/image_processing/image_demo.cpp.