Self supervised learning improves imaging flow cytometry with fewer annotations
IFC-SSL-Bench: Benchmarking Self-Supervised Learning for Annotation-Efficient Multispectral Imaging Flow Cytometry
Computer Vision and Pattern Recognition
Summary
Imaging flow cytometry creates large and complex biological images but requires experts to label them, which is slow and expensive. The authors show that self-supervised learning, a way for computers to learn from unlabeled images, can nearly match fully supervised methods using only a small fraction of labeled data. They tested several learning methods and network designs on different datasets, finding the best setups for tasks like identifying cell shapes and immune cell interactions. Their work offers guidelines for efficiently deploying these techniques and provides a benchmark to help others compare future methods.
What this means in practice
- •For biotechnology data teams: Improve cell image classification accuracy using self-supervised learning with minimal expert annotations for high-throughput biological experiments.
- •For medical imaging engineers: Design imaging analysis pipelines that reduce annotation costs by applying SSL methods optimized for complex cellular structure classification.
Authors
Angelo Jovin Yamachui Sitcheu, Roxana-Teodora Fetescu, Thomas Lautenschlager, Nils Friederich, Linus Schilpp, Stephan Allgeier, Luca Deininger, Daniel Hübschmann, Ralf Mikut
Abstract
Imaging Flow Cytometry (IFC) generates large multispectral datasets, yet the need for expert annotations bottlenecks the development of supervised deep learning models. While Self-Supervised Learning (SSL) offers a solution, its application to IFC remains fragmented and lacks systematic benchmarking. To address this, we present IFC-SSL-Bench, a modular framework for evaluating SSL in IFC. We benchmark six SSL approaches across four backbones on two heterogeneous datasets. By assessing downstream classification performance across annotation regimes of the training sets (1%-100%), we quantify annotation efficiency and identify optimal configurations for high-dimensional biological imaging. SSL matches the performance of fully supervised baselines trained on 100% of annotated data while using as little as 5% of annotations for morphological classification, with advantages in low-data regimes below 5%. Across both datasets, the contrastive approach MoCo-v2 paired with ConvNeXt backbones emerged as the top-performing configuration, with ConvNeXtV1-T best for immunological synapse classification and ConvNeXtV2-T for morphological classification. We identify a clear pattern in transfer dynamics: ImageNet initialization benefits morphological classification, consistent with its visual similarity to natural images, whereas random initialization better captures the more domain-specific immunological synapse classification. SSL elevates class-specific F1-scores by over 35 percentage points for structurally complex classes at 1% of annotated data. We synthesize these insights into a unified set of guidelines for deploying self-supervised workflows in high-throughput IFC, and release the benchmark as a resource for future work.