Scaling Channel-Invariant Self-Supervised Learning
Keywords: Self Supervised Learning, Vision Transformer, microscopy, channel adaptive
Abstract: Recent advances in self-supervised pre-training of foundation models for natural images have made them a popular choice for various visual systems and applications. Self-supervised strategies have also shown promise in non-RGB scientific imaging domains such as in biology, medical and satellite imagery, but their broader application is hampered by heterogeneity in channel composition and semantics between relevant datasets: two datasets may contain different numbers of channels, and these may reveal distinct aspects of an object or scene. Recent works on channel-invariant strategies report substantial advantages for those that account for variable channel compositions without sacrificing the ability to jointly encode channels; yet, how these strategies behave at scale remains unclear. We here show that, surprisingly, trained across large-scale microscopy datasets, independent-encoding of channels consistently outperforms joint-encoding methods by a substantial margin. We explore this result along an extensive set of experiments and open-source a new general purpose feature extractor for fluorescent microscopy images, DINO BoC, that sets a new state-of-the-art across challenging benchmarks, including generalization to out-of-distribution tasks and unseen channel combinations at test time.
Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 7797
Loading