Conclusions  

I introduced the LSJIMP method, a general least-squares inversion strategy to solve the problem of how to combine information from multiple and primary images while simultaneously suppressing crosstalk noise on each image. Successful tests on real 2-D and 3-D marine seismic data confirm that LSJIMP holds great promise as a novel and useful tool in the quest for a better seismic image. While my particular implementation of LSJIMP uses a fast, but relatively inaccurate imaging methodology, the continued trend toward inexpensive parallel computing implies that we may soon have the ability to use more accurate (and expensive) prestack imaging operators in LSJIMP.