how to split image and fill different color

Question

As this picture, how to use matlab code to split it into different parts, and then fill color in it? In addition, how to set gradient color in the second code ???

Here is the picture segmentation code:

clc
rgb=imread('sample1.bmp');
bw=im2bw(rgb2gray(rgb),.8);
bw=medfilt2(bw);
planes=bwareaopen(bw,800);
D=mat2gray(bwdist(imcomplement(planes)));
stats=regionprops(D>.8,'Centroid');
planes_centroid=cat(1,stats.Centroid);
planes_mask=false(size(bw));     
planes_mask(sub2ind(size(bw),round(planes_centroid(:,2)),...
        round(planes_centroid(:,1))))=1;
M=imimposemin(imcomplement(D),planes_mask);
L=watershed(M);
r=L & planes;
stats=regionprops(r,'BoundingBox','Centroid')
bb=cat(1,stats.BoundingBox);
c=cat(1,stats.Centroid);
figure,imshow(planes)
hold on
for i=1:length(stats)
   rectangle('Position',bb(i,:),'EdgeColor','b')
   plot(c(i,1),c(i,2),'r*')
   text(c(i,1)-5,c(i,2)-10,num2str(i))
end
%second code
clc;clf;close all;clear all;
color=cell(4,1);
for i=1:4
input=imread(['heartspline2_4_',num2str(i)],'bmp');  
figure,imshow(input);
BW=im2bw(input,graythresh(input));
[B,L]=bwboundaries(BW,'noholes');
for k=1:length(B)
   boundary=B{k};
   ind=size(boundary(:,2));
   plot(boundary(:,2),boundary(:,1),'k','LineWidth',2);
   hold on;
   axis off;
   if (k==1)
       patch(boundary(:,2),boundary(:,1),'w');
   else
       patch(boundary(:,2),boundary(:,1),???);
   end
end
saveas(gca,['y_','heartspline2_4_',num2str(i)],'bmp')
close(gcf)
end

Answer 1

You can use bwlabel to assign different index to each image region:

img = imread('http://i.stack.imgur.com/F1Iya.jpg');  %// read image
bw = img(:,:,1) > 128;  %// convert to binary mask
lb = bwlabel(bw,4);  %// extract distinct regions

The result:

figure; imshow(lb, [], 'border', 'tight'); colormap(rand(256,3));

---

If you want a gradient effect to the colors, you can

[x y] = meshgrid(linspace(0,1,size(bw,2)), linspace(0,1,size(bw,1)));
rand('seed',543310);
rgb_lb = ind2rgb(lb, rand(max(lb(:)+1),3)); %// convert to RGB color image
gx = x; 
gx(lb==1)=1;  %// use the horizontal gradient
gx = gx./max(gx(:));

Apply the gradient:

rgb_lb = bsxfun(@times, rgb_lb, gx);

The result: