DL-SMLM: a biological imaging dataset containing paired widefield and SMLM super-resolution images

DL-SMLM is a biological image dataset provided for deep-learning super-resolution microscopy, specially methods enabled by single-molecule localization microscopy (SMLM). DL-SMLM provides aligned low-resolution (LR) and high-resolution (HR) image pairs of subcellular structures in fixed cells, acquired from total internal reflection fluorescence microscopy (TIRF) and SMLM, respectively. This dataset consists of five biological structures: CCPs (clathrin heavy chain), ER membrane (sec61β), ER lumen (KDEL), microtubules (MAP7), and mitochondria outer membrane (TOMM20). For each LR image data, we provide a raw image stack with 100 frames captured at 50 ms per frame, which allows users to generate images with different signal-to-noise ratios. For each SR image data, we provide a reconstruction image (8× magnification) and a raw localization data, which allows users to generate customized SMLM reconstructions with different amplifications. For more information about DL-SMLM, please contact us. Please contact the author by e-mail [e-mail].

DL-SMLM是一款专为深度学习超分辨率显微技术（deep-learning super-resolution microscopy）设计的生物图像数据集，尤其适配单分子定位显微术（single-molecule localization microscopy, SMLM）相关方法。该数据集提供了固定细胞内亚细胞结构的配准低分辨率（low-resolution, LR）与高分辨率（high-resolution, HR）图像对，分别通过全内反射荧光显微镜（total internal reflection fluorescence microscopy, TIRF）与SMLM采集得到。本数据集涵盖五种亚细胞结构：网格蛋白重链（clathrin heavy chain, CCPs）、内质网膜（sec61β, ER membrane）、内质网腔（KDEL, ER lumen）、微管（MAP7, microtubules）以及线粒体外膜（TOMM20, mitochondria outer membrane）。针对每一份低分辨率图像数据，我们提供了单帧曝光时长50 ms、共计100帧的原始图像栈，可支持用户生成不同信噪比的图像。针对每一份超分辨率（super-resolution, SR）图像数据，我们提供了8倍放大的重建图像与原始定位数据，可支持用户生成不同放大倍率的定制化SMLM重建结果。如需了解DL-SMLM的更多相关信息，请通过邮箱[e-mail]联系作者。

搜集汇总

背景与挑战

背景概述

DL-SMLM是一个用于深度学习超分辨率显微镜的生物图像数据集，包含配对的低分辨率和高分辨率图像，覆盖五种生物结构（CCPs、ER膜、ER腔、微管和线粒体外膜）。数据集提供原始图像堆栈和重建图像，适用于不同信噪比和放大倍数的自定义重建。