DOC: transparency in use of SAM2 vs. SAM2.1 in reproducing manuscript results

[adamwitmer]

For manuscript associated with this code repository (la-2026-00306e), both SAM2 and SAM2.1 were used to generate results included there. There are performance differences between these two model weight parameters and so results from running the workflow may differ when using SAM2 vs. SAM2.1. Specifically, all of the results except those contained in the following list of tables/figures were generated using SAM2:

• Table S3: average precision values for all testing configurations
• Figure S22: (a) plot of the number of detected bubbles vs. image resolution and (b) the segmentation mask from sam2.1 detection
• Figure S23: the segmentation masks and filtered contour outputs from detecting irregularly shaped emulsions
• Figure S24: the PR curves associated with the output results shown in Table S3.

[tylerjereddy]

Sounds like something you should add to the README perhaps.

[tylerjereddy]

(and, in theory, it should be disclosed in the manuscript, which can cause quite some chaos to change once published)