Improve foreground detection

[camilolaiton]

At the moment, we are using parameters that were calculated for a set of SmartSPIM datasets to be able to detect foreground based on image intensity from the microscope. This was a quick fix to be able to process as quickly as possible imaged datasets without worrying too much about image quality at that moment due to further experimentation.

However, these parameters will not fit when the microscope/inmunolabeling/clearing is updated. We need to find a dynamic way to get the foreground threshold for each dataset. This will improve image quality after filtering with wavelet decomposition.