drift of average value of noisy realizations

[jeanollion]

Hello,
I have notice what seems to be a drift in the average values of the different images of a given FOV.
For a given FOV f, let GT_f be the average of the 400 noisy images I_f_i the drift in average value is D_f_i = | mean(GT_f) - mean(I_f_i) |
For channel 1, among all FOV: max_f(max_i(D_f_i)) = 17.8 and mean_f(mean_i(D_f_i))=1.14 when the average signal is 138.8 .
This induces a bias in computation of GT and makes the noise non-centered (I provide below the mean values of the noise N_f=I_f_249 - GT_f for each FOV of channel 1 ).
I'm working with the raw images.

In my case this impacts metrics of denoising methods.
Have you observed this issues, and do you have a standard correction for this ?
Thanks in advance.
Jean Ollion

noise mean for each FOV: [-0.8317243, 1.2392317, -4.208956, -3.358546, -4.2604985, 3.303461, -1.4710585, -0.2814204, 0.2828135, -0.449574, 2.147688, -1.1022385, -0.1850455, 0.5989662, 0.9377363, 0.5326594, 0.6375745, 1.7398707, 0.7471948, 0.2714147, 1.9833765, 0.19572908, 2.6688302, -2.2421582, -0.28200242, 0.08040069, 1.4105449, 1.1168568, -1.3297158, 1.0243058, 0.95828694, 1.2252651, -1.2509149, -0.24397486, 2.5610054, 1.8038473, 1.2389781, 0.62126774, 1.1971412, 1.1800888, -0.48276165, -2.0858142, -2.523765, 0.78499997, 3.0807197, 0.26181412, -1.2044606, 1.5227144, 0.14979905, 1.0882232, 0.13115424, -0.071832456, -0.15642868, -0.43484783, -0.6926114, -0.9822405, 1.15292, 0.45529222, 0.89752686, 0.009188991, -0.08303383, -2.2161212, 0.121987365, 0.3015615, 0.15540993, -4.771824, 0.23881689, -4.0210414, -0.7691488, -0.28321892, -1.5681504, -0.37152502, 2.7180696, 1.4099203, -2.3630157, -2.3754416, -1.2403578, 3.026292, -0.77643365, -2.202843, 2.1636384, 2.2502825, 2.3571348, 1.0598212, 0.017560417, -0.40768188, -0.6890089, -3.614546, 1.155397, 0.96262175, -0.68197405, 2.822749, -1.321134, -2.2414024, 0.031902194, -1.0697134, -0.20994323, 0.6897445, 1.9160875, -1.3230332, 3.8062916, 2.4175704, -0.46756917, 1.4648752, -0.018905409, -1.3221332, -0.31314403, 0.6417094, 1.1749882, 1.0345253, -0.87445986, 1.1762475, 3.9479492, -0.4438193, 0.064207785, -0.023243299, 0.07688828, 0.74707925, 0.8759693, -0.26784295] 
 mean=0.1133490651845932 std=1.6563199758529663