The study presented in this paper concerns an interactive fusion system that uses the fuzzy set theory to detect regions in three-dimensional seismic images. To achieve detection of regions in 3D seismic images, attributes extracted from images are fused using geophysicist interpreter knowledge by means of a fuzzy rule-based classifier. The original contribution of this work lies in the means proposed to the end-user for tuning the fuzzy membership functions in a two-dimensional universe for particular 2D reference image sections. The proposed graphic user interface allows to obtain a better region detection compared with the detection obtained without fusion. Moreover, a confidence index, based on information theory concepts, is introduced. This index is based on a coefficient of attribute influence and provides some elucidation on how the fusion results have been obtained.