dice

Sørensen-Dice similarity coefficient for image segmentation

Syntax

similarity = dice(BW1,BW2)

similarity = dice(L1,L2)

similarity = dice(C1,C2)

Description

similarity = dice(BW1,BW2) computes the Sørensen-Dice similarity coefficient between binary images BW1 and BW2.

similarity = dice(L1,L2) computes the Dice index for each label in label images L1 and L2.

similarity = dice(C1,C2) computes the Dice index for each category in categorical images C1 and C2.

Examples

Compute Dice Similarity Coefficient for Binary Segmentation

Read an image with an object to segment. Convert the image to grayscale, and display the result.

A = imread('hands1.jpg');
I = im2gray(A);
figure
imshow(I)
title('Original Image')

Use active contours (snakes) to segment the hand.

mask = false(size(I));
mask(25:end-25,25:end-25) = true;
BW = activecontour(I, mask, 300);

Read in the ground truth segmentation.

BW_groundTruth = imread('hands1-mask.png');

Compute the Dice index of the active contours segmentation against the ground truth.

similarity = dice(BW, BW_groundTruth);

Display the masks on top of each other. Colors indicate differences in the masks.

figure
imshowpair(BW, BW_groundTruth)
title(['Dice Index = ' num2str(similarity)])

Compute Dice Similarity Coefficient for Multi-Region Segmentation

This example shows how to segment an image into multiple regions. The example then computes the Dice similarity coefficient for each region.

Read an image with several regions to segment.

RGB = imread('yellowlily.jpg');

Create scribbles for three regions that distinguish their typical color characteristics. The first region classifies the yellow flower. The second region classifies the green stem and leaves. The last region classifies the brown dirt in two separate patches of the image. Regions are specified by a 4-element vector, whose elements indicate the x- and y-coordinate of the upper left corner of the ROI, the width of the ROI, and the height of the ROI.

region1 = [350 700 425 120]; % [x y w h] format
BW1 = false(size(RGB,1),size(RGB,2));
BW1(region1(2):region1(2)+region1(4),region1(1):region1(1)+region1(3)) = true;

region2 = [800 1124 120 230];
BW2 = false(size(RGB,1),size(RGB,2));
BW2(region2(2):region2(2)+region2(4),region2(1):region2(1)+region2(3)) = true;

region3 = [20 1320 480 200; 1010 290 180 240]; 
BW3 = false(size(RGB,1),size(RGB,2));    
BW3(region3(1,2):region3(1,2)+region3(1,4),region3(1,1):region3(1,1)+region3(1,3)) = true;
BW3(region3(2,2):region3(2,2)+region3(2,4),region3(2,1):region3(2,1)+region3(2,3)) = true;

Display the seed regions on top of the image.

imshow(RGB)
hold on
visboundaries(BW1,'Color','r');
visboundaries(BW2,'Color','g');
visboundaries(BW3,'Color','b');
title('Seed Regions')

Segment the image into three regions using geodesic distance-based color segmentation.

L = imseggeodesic(RGB,BW1,BW2,BW3,'AdaptiveChannelWeighting',true);

Load a ground truth segmentation of the image.

L_groundTruth = double(imread('yellowlily-segmented.png'));

Visually compare the segmentation results with the ground truth.

figure
montage({label2rgb(L),label2rgb(L_groundTruth)})
title('Comparison of Segmentation Results (Left) and Ground Truth (Right)')

Compute the Dice similarity index for each segmented region. The Dice similarity index is noticeably smaller for the second region. This result is consistent with the visual comparison of the segmentation results, which erroneously classifies the dirt in the lower right corner of the image as leaves.

similarity = dice(L, L_groundTruth)

similarity = 3×1

    0.9396
    0.7247
    0.9139

Input Arguments

BW1 — First binary image

logical array

First binary image, specified as a logical array of any dimension.

Data Types: logical

BW2 — Second binary image

logical array

Second binary image, specified as a logical array of the same size as BW1.

Data Types: logical

L1 — First label image

array of nonnegative integers

First label image, specified as an array of nonnegative integers, of any dimension.

Data Types: double

L2 — Second label image

array of nonnegative integers

Second label image, specified as an array of nonnegative integers, of the same size as L1.

Data Types: double

C1 — First categorical image

categorical array

First categorical image, specified as a categorical array of any dimension.

Data Types: category

C2 — Second categorical image

categorical array

Second categorical image, specified as a categorical array of the same size as C1.

Data Types: category

Output Arguments

similarity — Dice similarity coefficient

numeric scalar | numeric vector

Dice similarity coefficient, returned as a numeric scalar or numeric vector with values in the range [0, 1]. A similarity of 1 means that the segmentations in the two images are a perfect match. If the input arrays are:

• binary images, similarity is a scalar.

• label images, similarity is a vector, where the first coefficient is the Dice index for label 1, the second coefficient is the Dice index for label 2, and so on.

• categorical images, similarity is a vector, where the first coefficient is the Dice index for the first category, the second coefficient is the Dice index for the second category, and so on.

Data Types: double

More About

Dice Similarity Coefficient

The Dice similarity coefficient of two sets A and B is expressed as:

dice(A,B) = 2 * | intersection(A,B) | / ( | A | + | B | )

where |A| represents the cardinal of set A. The Dice index can also be expressed in terms of true positives (TP), false positives (FP) and false negatives (FN) as:

dice(A,B) = 2 * TP / ( 2 * TP + FP + FN)

The Dice index is related to the Jaccard index according to:

dice(A,B) = 2 * jaccard(A,B) / (1 + jaccard(A,B) )