Given some functional $E$ on a convex set $\Omega$, the Bregman divergence $D_E$ (of some convex function $E$) is defined at a point $p$ as the difference between its value at that point and its first order Taylor expansion around some other point $q$ :
$$D_E(p,q):=E(p)-E(q)-\langle \nabla E(q) , p-q\rangle. $$
For some strictly convex function $\phi$ the proximal map according to this Bregman divergence is
$$ \text{Prox}^{D_E}_\phi(p)=\text{argmin}_{q\in \mathcal{D}(\phi)}D_E(q,p)+\phi(q).$$
How can iterative Bregman projections be used for finding the infimum of $\phi$?