Connections between multitarget intensity filters and medical imaging

Abstract

The discussion of multitarget intensity tracking filters is greatly enlivened by exploring the close connections that exist between these filters and two medical imaging methods whose clinical use is well established. These connections arise because nonhomogeneous Poisson point processes (PPPs) play a central role in both applications as models of the spatial variability of the crucial quantity of interest. In tracking applications this quantity is the expected target density per unit state space, while in medical imaging applications it is the density per unit volume of radioisotopes absorbed by the body. Understanding the many connections between these fields is potentially beneficial to both; however, the immediate payoff is in the tracking application - the connection significantly clarifies the interpretation and meaning of the multitarget state model, and it validates the multisensor intensity filter as an averaging filter.