Convergence of the Galerkin Approximation to the Cardiac Bidomain Problem