Illustration of the iterative process used to determine the brightness threshold for constructing the segmentation map. We start by smoothing out the image data with a 10-pixel-wide Gaussian. Then, at each step, the brightness threshold is decreased incrementally, and the impact is assessed by examining (1) the relative increase in the area enclosed by the threshold (y-axis) and (2) the ratio of the threshold value to the mean brightness within the enclosed area (x-axis). The threshold is deemed sufficiently low when the target object can be distinguished from other objects in the image, which is determined by whether the cutoff curve (gray) has been crossed in this parameter space. The optimal shape of the cutoff curve was established through manual evaluation of a randomly selected subset of the sample.