CHANVESE

Active contour without edges.

Syntax
= (, , _, , )

Description
= (, , _, , ) performs active contour without edges on grayscale,
binary, or color image with initial phase .

Variable specifies the algorithm of active contour. Its default value is .

Value Meaning
Chan-Vese method (active contour without edge)
Chan-Vese method (active contour without edge for vector image)
Chan-Vese method (A Multiphase Level Set active contour)

Here is created by either the user or the built-in mask creation function. In order to define a user , the user
should make sure that the dimensions of match those of the input image , i.e. , 1 = , 1 and
, 2 = (, 2).

For = , only the top layer is used, i.e. , 3 = 1.

But for = , two layers is used, i.e. , 3 = 2.

For built-in mask creation function, keywords and their meanings are listed as follows. Depending on the value of
, the may have one layer or two layers.

Value Meaning
Creating a small circular mask with r = 9
Creating a medium circular mask
Creating a large circular mask
Creating a mask with small holes all over the mask
+ Creating a whole mask on one layer and another small mask on the other layer

_ is the iteration number for algorithm.

is the same variable defined in the paper, while 1 and 2 in the paper are simply defined to 1 . The
default value of is 0.2.

is the output segmentations gotten by the Chan-Vese active contour. is a binary image
for = and it is a labeled matrix for different partitions of the original image for =

.

1 By Yue Wu @ Tufts ECE

Class Support
When using a user defined , the can only be logical. Input image can be double, binary, gray and RGB.

Examples

Example 1: Creating Customerlized Mask in CHANVESE

Find the segmentation for a gray image with chan and customerlized .

I = imread('brain.jpg'); % read input image

m = zeros(size(I,1),size(I,2)); % initialize the dimensions of the mask
m(20:120,20:120) = 1; % create the mask
seg = chenvese(I,m,500,0.1,'chan'); % get segmentation

2 By Yue Wu @ Tufts ECE

Example 2: Active contour on noisy image

Show the strong anti-noise ability of active contours without edge for vector image.

First create a noisy image from Matlab. Two types of noises have been added in the original image.

P = imread('anti-mass.jpg'); % Get original image

I = P;
I(:,:,1) = imnoise(I(:,:,1),'speckle'); % add noise on component 1
I(:,:,2) = imnoise(I(:,:,2),'salt & pepper',0.8); % add noise on component 2
figure(),subplot(1,2,1),imshow(P),title('original image'); % show original & noisy image
subplot(1,2,2),imshow(I),title('original image with two components adding noise')

Method chan is not good enough for this noisy image. Its result is showed as follows.

seg = chenvese(I,'large',300,0.02,'chan');

3 By Yue Wu @ Tufts ECE

Use method vector i.e. Chan-Vese algorithm for vector image.

seg = chenvese(I,'large',300,0.02,'vector');

As we know, active contour is sensitive to its initial positions. Thus a different mask may lead to a different
segmentation. Here built-in = is used and it leads to better and faster result.

seg = chenvese(I,'whole',200,0.02,'vector');

4 By Yue Wu @ Tufts ECE

Example 3: Multiphase Active Contours

Get the segmentation via multiphase algorithm. For two phases, we can at most distinguish four partitions, namely four
colors.

First, we apply this algorithm on a synthesized image.

I = imread('4colors.jpg');
seg = chenvese(I,'whole',400,0.1,'multiphase');

5 By Yue Wu @ Tufts ECE

Now, we apply this algorithm to a real image.

I = imread('flowers.jpg');
seg = chenvese(I,'whole',800,0.2,'multiphase');

Reference
1. Chan, T. F., & Vese, L. A. (2001). Active contours without edges. IEEE Transactions on Image Processing, 10(2),
266277.
2. Chan, T.F., & Sandberg Y. B(2000). Active contours without edges for Vectorvalued Image. Journal of Visual
Communication and Image Representation 11, 130141 (2000)
3. Chan, T. F., & Vese, L. A. (2002). A Multiphase level set framework for image segmentation using the Mumford
and Shah model. International Journal of Computer Vision 50(3), 271293, (2002)
4. Kass, M., Witkin, A., & Terzopoulos, D. (1988). Snakes: Active contour models. International Journal of
Computer Vision, 1(4), 321331.
5. J. A. Sethian, Level Set Methods and Fast Marching Methods: Evolving Interfaces in Computational Geometry,
Fluid Mechanics, Computer Vision, and Materials Science (Cambridge ... on Applied and Computational
Mathematics), Cambridge University Press; 2 edition (1999)
6. Matlab Help, Mathworks Inc.
7. Shawn Lankton, Active Contour Matlab Code Demo, http://www.shawnlankton.com/?s=active+contour

6 By Yue Wu @ Tufts ECE