Image denoising for fluorescence microscopy by self-supervised transfer learning

When using fluorescent microscopy to study cellular dynamics, trade-off typically has to be made between light exposure and quality of recorded image to balance phototoxicity and image signal-to-noise ratio. Image denoising is an important tool for retrieving information from dim live cell images. Recently, deep learning based image denoising is becoming the leading method because of its promising denoising performance achieved by leveraging available prior knowledge about noise model and samples at hand. However, the practical application of this method has seen challenges because of the requirement of task relevant big training data. In this work, we show the approach of combining self-supervised learning with transfer learning to address the above challenge. We demonstrate the application of it in subcellular fluorescent imaging, where the light exposure dose can be significantly reduced and the spatial resolution is well restored.