Abstract
The fuzzy set theory is applied to the Signal Detection Theory to develop a new approach called the Fuzzy Signal Detection Theory (FSDT). First, linguistic variables are provided for the observer to describe his belief about the presence of a signal. Then, based on the assumptions that the judgment criterion (X c ) and sensitivity (d′) are both fuzzy, we use fuzzy integration to derive their membership functions, and thus give a more reasonable description of the signal detection mechanism. Further, a new index called the stability is proposed to investigate the consistency property of human judgment. It can be used as a selection criterion for signal detection personnels. Finally, some defuzzification operations are performed to put these results into use, and to develop a new tool for system comparison, i.e. the Fuzzy Receiver Operating Characteristic (FROC) curve. The fuzzy set approach is found to be more efficient and effective than the traditional ROC curve. © 1995.