Deep neural networks have enabled astonishing transformations from low-resolution (LR) to super-resolved images. However, whether, and under what imaging conditions, such deep-learning models outperform super-resolution (SR) microscopy is poorly explored. Here, using multimodality structured illumination microscopy (SIM), we first provide an extensive dataset of LR–SR image pairs and evaluate the deep-learning SR models in terms of structural complexity, signal-to-noise ratio and upscaling factor. Second, we devise the deep Fourier channel attention network (DFCAN), which leverages the frequency content difference across distinct features to learn precise hierarchical representations of high-frequency information about diverse biological structures. Third, we show that DFCAN’s Fourier domain focalization enables robust reconstruction of SIM images under low signal-to-noise ratio conditions. We demonstrate that DFCAN achieves comparable image quality to SIM over a tenfold longer duration in multicolor live-cell imaging experiments, which reveal the detailed structures of mitochondrial cristae and nucleoids and the interaction dynamics of organelles and cytoskeleton.

深度神经网络已经实现了从低分辨率 (LR) 到超分辨率图像的惊人转换。然而，关于这种深度学习模型是否以及在何种成像条件下优于超分辨率 (SR) 显微镜的研究却很少。在这里，使用多模态结构照明显微镜 (SIM)，我们首先提供了一个广泛的 LR-SR 图像对数据集，并根据结构复杂性、信噪比和放大因子评估深度学习 SR 模型。其次，我们设计了深度傅里叶通道注意力网络 (DFCAN)，它利用不同特征之间的频率内容差异来学习有关不同生物结构的高频信息的精确分层表示。第三，我们表明，DFCAN 的傅里叶域聚焦能够在低信噪比条件下稳健地重建 SIM 图像。我们证明，在多色活细胞成像实验中，DFCAN 在十倍长的时间内实现了与 SIM 相当的图像质量，这揭示了线粒体嵴和类核的详细结构以及细胞器和细胞骨架的相互作用动力学。