Deep learning enhanced fluorescence fluctuations super-resolution microscopy

Fluorescence fluctuations super-resolution microscopy (FF-SRM) is a powerful tool in imaging and monitoring of biological subcellular structures and dynamics in cells. A variety of image reconstruction algorithms have been developed for FF-SRM. In order to obtain high spatiotemporal resolution, a U-Net-based deep learning method for super-resolution imaging of fluorescence fluctuations was developed. With just 20 frames, super-resolution images reconstructed using U-Net model could be comparable to those reconstructed using VeSRRF algorithm with several hundred frames, demonstrating its capability of advancing imaging capabilities.