Updating segmentation models with more data generally improves performance, but what is the impact on the derived metrics? How stable are these metrics along model lifetime? Naga Karthik, Jan Valošek, @sandbedard.bsky.social and collaborators address this question in this latest publication 👇
