Discrimination of image signals on a background of bimodal Gaussoid pulsed noise, optimal by the combined-limitation criterion

Самойлин Е.А. Оптимальное по критерию объединенных ограничений выделение сигналов изображений на фоне бимодальных гауссовоподобных импульсных помех // Оптический журнал. 2010. Т. 77. № 4. С. 26–35.

Samoilin E.A. Discrimination of image signals on a background of bimodal Gaussoid pulsed noise, optimal by the combined-limitation criterion [in Russian] // Opticheskii Zhurnal. 2010. V. 77. № 4. P. 26–35.

E. A. Samoĭlin, "Discrimination of image signals on a background of bimodal Gaussoid pulsed noise, optimal by the combined-limitation criterion," Journal of Optical Technology. 77(4), 245-252 (2010). https://doi.org/10.1364/JOT.77.000245

Algorithms have been developed for discriminating image signals on a background of bimodal Gaussoid pulse noise under conditions of complete a priori indeterminacy concerning the image-signal distribution. The algorithms are based on a new optimality criterion that combines the minimum of the unconditional skip probability with the maximum correct-detection probability at limited levels of unconditional false-alarm and correct-nondetection probabilities. The results of numerical studies of algorithms are presented that show that they are of higher quality than known nonoptimal procedures, as well as than the optimal algorithm under the condition of complete a priori signal indeterminacy.

The research was supported by the Presidential grants council (grant No. МК-29.2009.10).