Figure 2.

Methods to define a donor threshold in space and time. Applying a threshold to the donor image before division allows masking irrelevant pixels (background around the luminescent specimen) in the A/D ratio image. Various threshold methods (B–F) can be applied to the donor image (A). Each column represents a different time point. (A) Intensity of the donor for each time point. (B) Mask using a constant threshold all over the image time series. (C,D) Mask using a global threshold computed for each time point (median threshold of the area outlined in red, (C) or Otsu thresholding method, (D). (E,F) Mask using a local threshold computed for each time point, Phansalkar (E) or Chastagnier threshold (F), described in Figure 3. Note that computing a local threshold for each image over the time series allows to precisely define the luminescent specimen limits, all along the experiment.