Deep-learning based denoising and reconstruction of super-resolution structured illumination microscopy images

• DOI: 10.1364/prj.416437
• Publication Date: 2021-01-29T17:30:24Z
• Authors: Zafran Hussain Shah; Marcel Müller; Tung-Cheng Wang; Philip Maurice Sc...; Axel Schneider; Mark Schüttpelz; Thomas Huser; Wolfram Schenck
• Keywords: 0301 basic medicine; 03 medical and health sciences; 530

Super-resolution structured illumination microscopy (SR-SIM) provides an up to twofold enhanced spatial resolution of fluorescently labeled samples. The reconstruction high-quality SR-SIM images critically depends on patterned with high modulation contrast. Noisy raw image data (e.g., as a result low excitation power or exposure time), in artifacts. Here, we demonstrate deep-learning based denoising that results reconstructed images. A residual encoding–decoding convolutional neural network (RED-Net) was used successfully denoise computationally noisy We also the end-to-end and SIM into high-resolution Both methods prove be very robust against artifacts generalize well across various noise levels. combination computational subsequent via RED-Net shows performance during inference after training even if microscope settings change.