Generalizing Stochastic Resonance by the Transformation Method

For a decision rule whose decision region is suboptimal we show how to transform the decision statistic to recover optimal performance. The procedure simply amounts to transforming the decision statistic to yield a combined statistic/decision function which is optimal. The approach may be thought of as a generalization of the stochastic resonance phenomenon, which employs a random linear transformation, and hence should be widely applicable to practical problems.