You could use getframe to get a matrix representing the image
x = linspace(-2*pi,2*pi);
y1 = i*j*sin(x);
y2 = i*j*cos(x);
plot(x,y1,x,y2);
f=getframe(gcf);
im=frame2im(f);
Now you have a bunch of numbers to work with. If you are against using RGB2GRAY and IMBINARIZE you have to decide how you'd like to turn it into zeros and 1s...is it everything that's not exactly white?
imshow(~any(im<255,3))
Or maybe a slightly reduced threshold?
imshow(~any(im<200,3))
To turn these zeros and ones into pixels, you'll want to transform it to integers between one and 255.
outim = uint8(255*~any(im<200,3));
We can use interp2 to resize by interpolation, to do that we'll need a grid of values ranging over the size of the image, and we'll want 64 values in our final image. meshgrid is an easy way to make this kind of grid:
[xi,yi]=meshgrid(linspace(1,width(outim),64),linspace(1,height(outim),64));
Now we can interpolate, and use the nearest option which will make sure that everything stays binary.
outim=interp2(outim,xi,yi,'nearest');
