Millisecond focal plane telemetry is now becoming practical due to a new generation of near-IR detector arrays with sub-electron noise that are capable of kHz readout rates. Combining these data with those simultaneously available from the wavefront sensing system allows the possibility of self-consistently determining the optical aberrations (the cause of quasi-static speckles) and the planetary image. This approach may be especially advantageous for finding planets within about 3 λ/D of the star where differential imaging is ineffective. As shown in a recent article by the author (J. Opt. Soc. Am. A., 33, 712, 2016), one must account for unknown aberrations in several non-conjugate planes of the optical system, which, in turn, requires ability to computational propagate the field between these planes. These computations are likely to be difficult to implement and expensive. Here, a far more convenient alternative based on empirical Green's functions is provided. It is shown that the empirical Green's function (EGF), which accounts for all multi-planar, non-common path aberrations, and results in a much more tractable and highly parallel computational problem. It is also shown that the EGF can be generalized to treat polarization, resulting in the empirical Green's tensor (EGT).