Using pixel intensity as a self-regulating threshold for deterministic image sampling in Milano Retinex: the T-Rex algorithm

Abstract

Using pixel intensity as a selfregulating threshold for deterministic image sampling in Milano Retinex: the T-Rex algorithm,"

Point‐based spatial colour sampling in Milano‐Retinex: a survey

“…Thus, when applying swRSR on an image containing shadows, these are lightened, and the value of f 2 decreases while the values of f 0 , f 1 increase as well as the differences with the original image. [20] 11.14 Local-QBRIX [20] 18.15 T-Rex [20] 21.17 GREAT [20] 18.92 GRASS [20] 16.60…”

“…iv. T-Rex [20], i.e. a deterministic approach that computes the lightness at any target x by thresholding a function of the image intensity and of the spatial distance of the image pixels from x; one important feature of this algorithm is that the threshold is a self-regulating parameter, equal to the target intensity.…”

“…T-Rex is a deterministic Milano-Retinex implementation, firstly presented in [32] and recently extended in [20]. T-Rex is based on an intensity thresholding strategy.…”

“…While in [15,21,22,32,34,35,37], the measures f 0 , f 1 and f 2 were computed on the luminance of the input colour image, in [20] as well as in this work, the f i 's are estimated them channel by channel. This allows to better characterise the performance of the algorithms discussed here.…”

“…Sections 3-6 describe, respectively, RSR [13] and its variant Light-RSR [14], spatially weighted (swRSR) [15], and the exact population-based formulation RSR-P [16]. Section 7 describes the probabilistic approach quantile-based approach to RetIneX (QBRIX) [17], followed by spatio-temporal Retinex-like envelope with stochastic sampling (STRESS) [18] and spatio-temporal Retinex-like envelope with total variation (STRETV) [19] in Section 8 and T-Rex [20] in Section 9. Sections 10 and 11 explain the algorithms GREAT [21] and GRASS [22].…”