Looking at the algorithm in the GERMLINE paper, it appears GERMLINE demonstrates an algorithm for finding IBD between haplotypes, but it does not address inbreeding. In research on humans and commercial genealogy dbs Is there typically a post processing step to prevent overestimating IBD due to inbreeding? Or is that mostly a theoretical concern?

For example, looking at the image below:

Where a solid line between the dots represents IBD, just doing pairwise haplotype comparisons across individuals would cause you to compute high IBD values in S1, S3, and S5.