Generalization across animals and acquisition protocols

Determine whether the fully automatic 3D dendrite instance-segmentation pipeline generalizes reliably to other animals and acquisition protocols.

Background

The pipeline was evaluated using one control animal and one pilocarpine-induced epileptic animal, while the random-forest inter-slice linker and nnU-Net refinement model were trained using only 30 annotated slices from the control dataset. Although the authors consider transfer to epileptic tissue plausible, the limited number of animals, training slices, and acquisition settings leaves the method's broader generalization unresolved.

Establishing generalization is important for determining whether the modular pipeline can be used reliably across biological specimens and imaging conditions rather than only within the specific datasets and acquisition protocol used in the study.

References

These results come from one animal per condition, and the RF linker and nnU-Net were trained on only 30 annotated slices of Dataset A, so generalization to other animals and acquisition protocols remains to be established.

— A Fully Automatic Pipeline for 3D Dendrite Instance Segmentation in SBF-SEM  (2610.03332 - Zhuo et al., 2 Oct 2026) in Discussion and Conclusion

Our training data cover human, cat, and dog drivers. Generalization to unseen species is therefore evaluated only qualitatively on the source side, using drivers from the trained species, while unseen-species driving and arbitrary facial topologies remain unexplored.

— CrossPortrait: Cross-Species Portrait Animation via Adversarial Motion Alignment  (2610.17532 - Pan et al., 13 Oct 2026) in Section Limitations