adaptthresh

Adaptive image threshold using local first-order statistics

Syntax

T = adaptthresh(I)

T = adaptthresh(I,sensitivity)

T = adaptthresh(___,Name=Value)

Description

T = adaptthresh(I) calculates a locally adaptive threshold for 2-D grayscale image or 3-D grayscale volume I. The adaptthresh function chooses the threshold based on the local mean intensity (first-order statistics) in the neighborhood of each pixel. The threshold T can be used with the imbinarize function to convert the grayscale image to a binary image.

T = adaptthresh(I,sensitivity) calculates a locally adaptive threshold using a sensitivity factor towards thresholding more pixels as foreground.

T = adaptthresh(___,Name=Value) calculates a locally adaptive threshold using name-value pairs to control aspects of the thresholding.

Examples

Find Threshold and Segment Bright Rice Grains from Dark Background

Read image into the workspace.

I = imread('rice.png');

Use adaptthresh to determine threshold to use in binarization operation.

T = adaptthresh(I, 0.4);

Convert image to binary image, specifying the threshold value.

BW = imbinarize(I,T);

Display the original image with the binary version, side-by-side.

figure
imshowpair(I, BW, 'montage')

Find Threshold and Segment Dark Text from Bright Background

Read image into the workspace.

I = imread('printedtext.png');

Using adaptthresh compute adaptive threshold and display the local threshold image. This represents an estimate of average background illumination.

T = adaptthresh(I,0.4,'ForegroundPolarity','dark');
figure
imshow(T)

Binarize image using locally adaptive threshold

BW = imbinarize(I,T);
figure
imshow(BW)

Calculate Threshold for 3-D Volume

Load 3-D volume into the workspace.

load mristack;
V = mristack;

Display the data.

figure
slice(double(V),size(V,2)/2,size(V,1)/2,size(V,3)/2)
colormap gray
shading interp

Calculate the threshold.

J = adaptthresh(V,'neigh',[3 3 3],'Fore','bright');

Display the threshold.

figure
slice(double(J),size(J,2)/2,size(J,1)/2,size(J,3)/2)
colormap gray
shading interp

Input Arguments

I — Grayscale image or volume

2-D numeric matrix | 3-D numeric array

Grayscale image or volume, specified as a 2-D numeric matrix or 3-D numeric array.

The adaptthresh function expects images of data type double and single to have values in the range [0, 1]. If I has values outside the range [0, 1], then you can rescale values to the expected range by using the rescale function.

If the image contains Inf or NaN values, the behavior of adaptthresh is undefined. Propagation of Inf or NaN values might not be localized to the neighborhood around Inf or NaN pixels.

Data Types: single | double | int8 | int16 | int32 | uint8 | uint16 | uint32

sensitivity — Determine which pixels get thresholded as foreground pixels

0.5 (default) | number in the range [0, 1]

Determine which pixels get thresholded as foreground pixels, specified as a number in the range [0, 1]. High sensitivity values lead to thresholding more pixels as foreground, at the risk of including some background pixels.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Example: T = adaptthresh(I,0.4,ForegroundPolarity="dark"); specifies that the foreground is darker than the background.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: T = adaptthresh(I,0.4,"ForegroundPolarity","dark"); specifies that the foreground is darker than the background.

NeighborhoodSize — Size of neighborhood used to compute local statistic around each pixel

2*floor(size(I)/16)+1 (default) | positive odd integer | 2-element vector of positive odd integers

Size of neighborhood used to compute local statistic around each pixel, specified as a positive odd integer or a 2-element vector of positive odd integers.

Data Types: single | double | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32 | uint64

ForegroundPolarity — Determine which pixels are considered foreground pixels

"bright" (default) | "dark"

Determine which pixels are considered foreground pixels, specified using one of the following:

Value	Meaning
"bright"	The foreground is brighter than the background.
"dark"	The foreground is darker than the background.

Data Types: char | string

Statistic — Statistic used to compute local threshold

"mean" (default) | "median" | "gaussian"

Statistic used to compute local threshold at each pixel, specified as one of the following:

Value	Meaning
"mean"	The local mean intensity in the neighborhood. This technique is also called Bradley’s method [1].
"median"	The local median in the neighborhood. Computation of this statistic can be slow. Consider using a smaller neighborhood size to obtain faster results.
"gaussian"	The Gaussian weighted mean in the neighborhood.

Data Types: char | string

Output Arguments

T — Normalized intensity values

numeric matrix | numeric array

Normalized intensity values, returned as a numeric matrix or numeric array of the same size as the input image or volume, I. Values are clipped to the range [0, 1].

Data Types: double

References

[1] Bradley, D., G. Roth, "Adapting Thresholding Using the Integral Image," Journal of Graphics Tools. Vol. 12, No. 2, 2007, pp.13–21.

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

• adaptthresh supports the generation of C code (requires MATLAB^® Coder™). Note that if you choose the generic MATLAB Host Computer target platform, adaptthresh generates code that uses a precompiled, platform-specific shared library. Use of a shared library preserves performance optimizations but limits the target platforms for which code can be generated. For more information, see Types of Code Generation Support in Image Processing Toolbox.

• The ForegroundPolarity and Statistic arguments must be compile-time constants.

GPU Code Generation

Generate CUDA® code for NVIDIA® GPUs using GPU Coder™.

Usage notes and limitations:

The ForegroundPolarity and Statistic arguments must be compile-time constants.

Thread-Based Environment

Run code in the background using MATLAB® backgroundPool or accelerate code with Parallel Computing Toolbox™ ThreadPool.

This function fully supports thread-based environments. For more information, see Run MATLAB Functions in Thread-Based Environment.

Version History

Introduced in R2016a

R2022b: Generate CUDA code using GPU Coder

adaptthresh now supports the generation of optimized CUDA^® code (requires GPU Coder™).

R2021b: Support for thread-based environments

adaptthresh now supports thread-based environments.