发表机构
Capicú Technologies（Capicú科技公司）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究评估了压缩的Cellpose-SAM在干细胞显微镜下的表现，提出按模态分层保留标准，发现混合W4/W8量化在6.76倍压缩下保持性能，而三元量化失败，强调下游保留评估的重要性。
AI 中文摘要
诱导多能干细胞（iPSC）培养日益依赖分割基础模型，但在实验室CPU和边缘硬件上的部署需要既高效又可审计的压缩方案。我们使用预定的保留标准对压缩后的Cellpose-SAM进行了面向部署的评估：对于每种成像模态，FP32平均变化的95%聚类自助法区间必须保持在固定的-0.02容差之上。在涵盖BBBC038细胞核、BBBC039 U2OS荧光和NIST iPSC图像（跨越不同密度状态）的分层176视野面板上，仅权重的W8A16在所有模态中保持了实例F1分数。一种基于敏感性引导的混合W4/W8方案，使用四个INT8例外，实现了权重存储减少6.76倍，且未观察到灾难性失败（0/176视野），在此样本量下与W8A16相当。相比之下，仅权重的三元量化实现了12.08倍压缩，但在169/176视野上灾难性失败。这些结果表明，压缩应通过按模态分层的下游保留来评估，而非单一数字的准确性，并为在受监管的干细胞成像中审计压缩基础模型建立了可复现的协议。
英文摘要
Induced pluripotent stem cell (iPSC) culture increasingly relies on segmentation foundation models, yet deployment on laboratory CPUs and edge hardware requires compression schemes that are both efficient and auditable. We present a deployment-oriented evaluation of compressed Cellpose-SAM using a pre-specified retention criterion: the 95% cluster-bootstrap interval of mean change from FP32 must remain above a fixed -0.02 margin for every imaging modality. On a stratified 176-field panel spanning BBBC038 nuclei, BBBC039 U2OS fluorescence, and NIST iPSC images across density regimes, weight-only W8A16 preserves instance F1 across all modalities. A sensitivity-guided mixed W4/W8 scheme, using four INT8 exceptions, achieves a 6.76x reduction in weight storage with no observed catastrophic failures (0/176 fields), matching W8A16 at this sample size. In contrast, ternary weight-only quantization achieves 12.08x compression but fails catastrophically on 169/176 fields. These results demonstrate that compression should be evaluated by modality-stratified downstream retention rather than single-number accuracy, and establish a reproducible protocol for auditing compressed foundation models in regulated stem-cell imaging.
CommentsPeer-reviewed and accepted at NeurIPS 2026 LXAI Workshop. Pending presentation/publication at JLXAIR
