Improve data intensity filter
When creating a dataset, there is an option to divide the data into the frames with high intensity (usable frames) and frames with low intensity (can be used as background). Now, this division is done by computing the variance of each blurred frame and creating its histogram. The frames that fall in the first bin are considered low intensity frames. But this varies on the number of bins one picks. The objective is to try to automatically find this division without any user input on the data.
We could pass the intensities vector to a Change Point Detector to try to find the best separation, https://cran.r-project.org/web/packages/cpm/index.html. This is not implemented in Python so it would have to be implemented.
Edited by Julia Garriga Ferrer