THE APPROXIMATE DISTRIBUTION OF THE MAXIMUM OF A SMOOTHED POISSON RANDOM FIELD
Daniel J. Rabinowitz, David Siegmund · 1997
The distribution of the maximum score statistic for detecting a signal of known shape, but unknown amplitude, location, and scale is discussed when the underlying noise process is a homogeneous Poisson process. The approximation is based on an exponential change of measure to evaluate asymptotically the expected number of local maxima of a random field.