PHOEBI: An Open-World Benchmark for Bacterial Identification in Phase-Contrast Microscopy

2026-06-22

AI Summary

This study addresses the limited generalization of existing computer vision methods to unseen multi-bacterial mixtures in phase-contrast microscopy and the absence of evaluation benchmarks for open-world scenarios. To this end, we introduce PHOEBI, a wet-lab dataset comprising 120,000 images spanning 40 combinations of six rod-shaped bacterial species, along with a leave-combination-out cross-validation protocol to simulate realistic open-world conditions. We propose a lightweight anchor decoder coupled with a frozen feature-pool sharing mechanism and linear probing to systematically evaluate the representational capabilities of thirteen encoder architectures. Experimental results demonstrate that conventional aggregators suffer substantial performance degradation on held-out combinations, with F1 scores dropping by 0.39–0.57, whereas our approach achieves significantly improved performance on unseen bacterial combinations, validating its effectiveness.

Abstract

Optical microscopy enables rapid, label-free imaging of live bacteria and is the standard instrument for species identification across clinical, environmental, and industrial microbiology. Yet field samples are routinely polymicrobial and may contain organisms that were never seen during system training, and no computer-vision benchmark tests multi-label species identification from phase-contrast microscopy (PCM) of such mixtures. We introduce Phase-contrast Optical bEnchmark for Bacterial Identification ($\textbf{PHOEBI}$), a wet-lab-prepared dataset of $120{,}000$ PCM images covering $40$ combinations of six rod-shaped species, paired with a leave-combinations-out (LCO) evaluation protocol that holds out entire species combinations to mirror the practical scenario of a model trained on catalogued mixtures that must generalise to unseen ones. On LCO, every gradient-trained per-image aggregator we test drops $0.39$ to $0.57$ F1 from the in-distribution to the held-out split, a systematic open-world recognition failure in the aggregator, not the visual representation. A linear probe of thirteen different encoders over the same features spreads only about six percentage points of F1 across general-purpose and biomedical pretraining objectives, confirming the representation is sound. We propose three lightweight $\textit{anchor-based}$ decoders that capture per-species presence geometrically over a shared frozen tile-feature pool, scoring $\textit{higher}$ on held-out combinations than on in-distribution validation.

Problem

• bacterial identification
• phase-contrast microscopy
• open-world recognition
• polymicrobial mixtures
• multi-label classification

Innovation

• open-world recognition
• phase-contrast microscopy
• bacterial identification
• leave-combinations-out
• anchor-based decoder