i have a process for doing it, but the complications are just too insane to consider.
if you let z = x + yi, square that,
w= (x^2 - y^2) + (2xy)i
then from the cartesian equation of the curve in a, get y in terms of x, so you can get w solely in terms of x.
then you have an Re(z) and an...