fibermetric

Enhance elongated or tubular structures in image using Frangi vesselness filter

Syntax

J = fibermetric(I)

J = fibermetric(I,thickness)

J = fibermetric(___,Name,Value)

Description

J = fibermetric(I) enhances elongated or tubular structures in the 2-D or 3-D grayscale image I using a Hessian-based multiscale Frangi vesselness filter. The image returned, J, contains the maximum response of the filter at a thickness that approximately matches the size of the tubular structure in the image.

J = fibermetric(I,thickness) specifies the thickness of the tubular structures to enhance.

J = fibermetric(___,Name,Value) uses name-value pair arguments to control different aspects of the filtering algorithm.

Examples

Find Threads Approximately Seven Pixels Thick

Read and display an image that contains tubular threads of varying thicknesses.

I = imread('threads.png');
imshow(I)

Create an enhanced version of the image that highlights threads that are seven pixels thick. Threads show up dark against a light background, therefore specify the object polarity as 'dark'. Display the enhanced image.

B = fibermetric(I,7,'ObjectPolarity','dark');
imshow(B)
title('Enhanced Tubular Structures 7 Pixels Thick')

Threshold the enhanced image to create a binary mask image containing the threads with the specified thickness.

BW = imbinarize(B);

Display the mask over the original image by using the labeloverlay function. The overlay has a blue tint where the mask is true (where threads have the specified thickness).

imshow(labeloverlay(I,BW))
title('Detected Tubular Structures 7 Pixels Thick')

Input Arguments

I — Image with elongated or tubular structures

2-D grayscale image | 3-D grayscale volume

Image with elongated or tubular structures, specified as 2-D grayscale image or 3-D grayscale volume.

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

thickness — Thickness of tubular structures

[4 6 8 10 12 14] (default) | positive integer | vector of positive integers

Thickness of tubular structures in pixels, specified as a positive integer or vector of positive integers.

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

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.

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

Example: J = fibermetric(I,'StructureSensitivity',15)

StructureSensitivity — Structure sensitivity

positive number

Structure sensitivity, specified as the comma-separated pair consisting of 'StructureSensitivity' and a positive number. The structure sensitivity is a threshold for differentiating the tubular structure from the background.

The default value depends on the data type of image I, and is calculated as 0.01*diff(getrangefromclass(I)). For example, the default threshold is 2.55 for images of data type uint8, and the default is 0.01 for images of data type double with pixel values in the range [0, 1].

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

ObjectPolarity — Polarity of tubular structures

'bright' (default) | 'dark'

Polarity of the tubular structures with the background, specified as the comma-separated pair consisting of 'ObjectPolarity' and one of the following values:

Value	Description
'bright'	Structure is brighter than the background.
'dark'	Structure is darker than the background.

Data Types: char | string

Output Arguments

J — Enhanced image

numeric array

Enhanced image, returned as a numeric array of the same size as the input image I. If the data type of I is double, then the data type of J is also double. Otherwise, the data type of J is single.

Data Types: single | double

Tips

• The fibermetric function does not perform segmentation. The function enhances an image to highlight structures and is typically used as a preprocessing step for segmentation.

References

[1] Frangi, Alejandro F., et al. Multiscale vessel enhancement filtering. Medical Image Computing and Computer-Assisted Intervention — MICCAI'98. Springer Berlin Heidelberg, 1998. pp. 130–137.

Extended Capabilities

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 R2017a

R2022b: Support for thread-based environments

fibermetric now supports thread-based environments.

R2018b: fibermetric calculates new default structural sensitivity

Starting in R2018b, fibermetric calculates the default value of the StructureSensitivity argument as 0.01*diff(getrangefromclass(I)). The change in the default value increases computational efficiency and reduces memory usage. These improvements also enable processing 3-D volumetric images.

Previous versions of fibermetric defined the default value of StructureSensitivity as half of the maximum of the Hessian norm of the image. If you want to reproduce the prior default value for 2-D images, then specify StructureSensitivity as 0.5*maxhessiannorm(double(I),thickness). The maxhessiannorm function does not support 3-D input.