How to filter image background

I used this code

mu=1;
nsamples=1000000;
L=1;
l=1;
a1=1;
a2=1;
x01=-4+8*rand(nsamples,1);
y01=-4+8*rand(nsamples,1);
r01=sqrt(x01.^2+y01.^2);
rm1=mu*r01.*exp(-r01.^2/a1^2)+r01;
A1=atan2(y01,x01);
xm1=rm1.*cos(A1);
ym1=rm1.*sin(A1);

x02=-4+8*rand(nsamples,1);
y02=-4+8*rand(nsamples,1);
r02=sqrt(x02.^2+y02.^2);
rm2=mu*r02.*exp(-r02.^2/a2^2)+r02;
A2=atan2(y02,x02);
xm2=rm2.*cos(A2)+2;
ym2=rm2.*sin(A2);

A=reshape([xm1,ym1],[],2);
B=reshape([xm2,ym2],[],2);
A1=A(:,[1]);
A2=A(:,[2]);
B1=B(:,[1]);
B2=B(:,[2]);
x=vertcat(A1,B1);
y=vertcat(A2,B2);
values=hist3([x,y],[300,300]);
image=mat2gray(values);
image2=imcomplement(image);
image3=imrotate(image2,90);
x=linspace(-5,5,10000);
y=linspace(-5,5,10000);
figure
imshow(flipud(image3))

to generate this image:

What I am trying to do is eliminate the background noise. I want the background to be white, starting just outside of the circles I generated. I tried changing r01 and r02 but this made the image look artificial. I also tried changing the scale of mat2gray but couldn't come up with the right values.

I know that there are some redundancies in my code and it isn't very elegant---part of this reason is to make it more clear for application (ie differentiation of x01 and x02, use of L,l,a1,a2). This should be irrelevent for the image manipulation