Minimum mean-square-error filter for pattern recognition with spatially disjoint signal and scene noise

Distortion-Invariant Pattern Recognition with Adaptive Correlation Filters

“…If an input scene contains a reference object embedded into a disjoint background (nonoverlapping model) and additive noise, the following correlation filters were derived: the generalized matched filter (GMF) maximizes the ratio of the expected value of the squared correlation peak to the average output variance (Javidi & Wang, 1994), the generalized phase-only filter (GPOF) maximizes the light efficiency (Kober et al, 2000), and the generalized optimum filter (GOF) maximizes the ratio of the expected value of the squared correlation peak to the average expected value of the output signal energy (POE) (Javidi & Wang, 1994). Other generalized filters were also introduced (Goudail & Réfrégier, 1997;Javidi et al, 1996;Réfrégier, 1999;Réfrégier et al, 1993;Towghi & Javidi, 2001). Conventional filters are sensitive to intensity signal degradations.…”

Correlation Pattern Recognition in Nonoverlapping Scene Using a Noisy Reference

“…Since the pioneering work by VanderLugt [1], correlation filters have been extensively studied for the purpose of pattern recognition [2][3][4][5][6][7][8][9][10][11][12][13][14][15]. Within the context of pattern recognition, detection and location estimation are two very important tasks.…”

“…The optimal filter (OF) [5] minimizes the probability of anomalous errors. Several filters have been derived for the nonoverlapping scene model [6,7,8,9,10]. The generalized matched filter [7] was derived by maximizing the ratio of the square of the expected value of the correlation peak to the average output variance.…”

Correlation Filters for Pattern Recognition Using a Noisy Reference

“…Since the pioneering work by VanderLugt [1], correlation filters have been extensively studied for the purpose of pattern recognition [2][3][4][5][6][7][8][9][10][11][12][13][14][15]. Within the context of pattern recognition, detection and location estimation are two very important tasks.…”

“…The optimal filter (OF) [5] minimizes the probability of anomalous errors. Several filters have been derived for the non-overlapping scene model [6,7,8,9,11]. The generalized matched filter [7] was derived by maximizing the ratio of the square of the expected value of the correlation peak to the average output variance.…”