Automatic FLIR resolution measurement using fuzzy logic and neural networks

Pacific-Sierra Research Corporation (PSR) proposes to utilize fuzzy logic to map Boundary Contour System (BCS)/Feature Contour System (FCS) output signtals to a "can discriminate"/"cannot discriminate" visual cognitive decision model. A set of FLIR images and corresponding human discrimination decision results will be utilized as training data for a fuzzy logic mapping network. Detection thresholds could be based on local statistical measures of BCS boundary strength or on more global statistical measures of FCS brightness levels inside and outside of the test bars. Which measure to employ depends upon reports from subjects regarding their detection strategy, as well as correlations between psychophysical detection thresholds and statistical variations in the BCS/FCS data representation. This will be a subject of the proposed research.