未明确命名

本研究使用了一个包含750张图像的小型数据集，用于训练基于卷积神经网络的超分辨率荧光显微镜模型。数据集来源于15个不同的视场，每张图像通过超分辨率径向波动方法生成。该数据集的创建旨在克服传统深度学习技术对大量图像的需求，通过新方法实现在小数据集上训练网络以达到超分辨率成像。此技术可应用于其他生物医学成像领域，如MRI和X射线成像，解决获取大型训练数据集的挑战。

This study utilizes a small dataset consisting of 750 images for training convolutional neural network-based super-resolution fluorescence microscopy models. The dataset is sourced from 15 distinct fields of view, with each image generated using the super-resolution radial fluctuations (SRRF) method. This dataset was developed to overcome the large-scale image data requirements of traditional deep learning technologies, enabling neural networks to be trained on small datasets to achieve super-resolution imaging through novel approaches. This technology can be applied to other biomedical imaging domains such as MRI and X-ray imaging, resolving the challenge of obtaining large-scale training datasets.