Anyone could explain the function of bwdist for me?

Hello, everyone. I learnt a example from help ,which was aim to use watershed function to segment the image.Here is the code

if true
  %  center1=-10;
center2=-center1;
dist=sqrt(3*(2*center1)^2);
radius=dist/2 * 1.4;
lims=[floor(center1-1.2*radius) ceil(center2+1.2*radius)];
[x,y,z]=meshgrid(lims(1):lims(2));
bw1=sqrt((x-center1).^2+(y-center1).^2+(z-center1).^2)<=radius;
bw2=sqrt((x-center1).^2+(y-center2).^2+(z-center2).^2)<=radius;
bw=bw1|bw2;
figure;
%extract isosurface from volume data 从容积数据里面提取等值曲面
isosurface(x,y,z,bw,0.5);
% 0.5 means isovalue; bw is the volume data; 
axis equal;
title('original');
set(gcf,'color','w');
%set:set graphics object properties;
%gcf means current figure handle 
% this sentence means set current graphics color to white
xlabel x, ylabel y,zlabel z
xlim(lims), ylim(lims), zlim(lims) %图像的边界
view(3)%set the default 3D view
%计算三维变换；
D=bwdist(~bw);%bwdist distance transform of binary image.
figure;
isosurface(x,y,z,D,radius/2);
axis equal;
title('距离等值变换');
set(gcf,'color','w');
xlabel x,ylabel y,zlabel z;
xlim(lims),ylim(lims),zlim(lims);
view(3);
camlight;%在camera坐标系里面 create or move light
lighting gouraud;% this method is used for curved surface
D=-D;
D(~bw)=-Inf; %force pixels do not belong to the image be -Inf
compute the watershed transform
L=watershed(D); %分水岭的算法，D为待分割的图像
figure;
isosurface(x,y,z,L==2,0.5);
isosurface(x,y,z,L==3,0.5);
axis equal;
title('分割对象');
set(gcf,'color','w');
xlabel x,ylabel y,zlabel z;
xlim(lims),ylim(lims),zlim(lims);
view(3),camlight,lighting gouraud

end
when I run this demo it was right as I expected but there is something that troubles me.

if true
  % D=bwdist(~bw)
end

why we need to compute the distance transform of complement? why not force the pixels which do not belong to this to be Inf.

Answer 1

[Image Analyst]

Because it's the distance to the nearest non-zero pixel, not the closest zero pixel. So if you want to find the distances within a white blob to the black background, you must invert the image to make the background white.