From the help section of that file,
% x,y are column vector where (x(i),y(i)) is a measured point
Their class are double.
They are column vectors of any length but must have the same length between x and y.
Demo
Here's a demo; see the inline comments to understand each step.
First, compute noisy circle coordinates for the demo.
% Create noisy circle coordinates
radius=3;
theta=linspace(0,2*pi,200);
theta(randi(numel(theta),1,190)) = []; % remove some data
x=radius*cos(theta);
y=radius*sin(theta);
% add noise
x = x + (rand(size(x))-0.5)*radius/4;
y = y + (rand(size(y))-0.5)*radius/4;
% Plot it
plot(x,y,'o')
axis equal
Do the fitting but you have to fill in the inputs. The reason I'm not telling you what the inputs should be is not to cause frustration but to get you to think about this. Again, the two inputs are column vectors of your x and y coordinates of each point you're fitting.
% Fit it
[xc,yc,R,a] = circfit(___, ___); % FILL IN THE INPUTS HERE
% [update: here's the solution: [xc,yc,R,a] = circfit(x, y);
Now you can use the outputs to compute the fitted circle and plot it.
% compute circle based on fit params
theta=linspace(0,2*pi,200);
xFit=R*cos(theta) + xc;
yFit=R*sin(theta) + yc;
% add fit
hold on
plot(xc,yc,'rx') % plot circle center
plot(xFit,yFit, 'r-')
