Single-Shot Autofocusing of Microscopy Images Using Deep Learning

• DOI: 10.1021/acsphotonics.0c01774
• Publication Date: 2021-01-22T03:20:19Z
• Authors: Yilin Luo; Luzhe Huang; Yair Rivenson; Aydogan Ozcan
• Keywords: FOS: Computer and information sciences; 0301 basic medicine; Computer Science - Machine Learning; brightfield microscopy modalities; Deep-R; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition; FOS: Physical sciences; Microscopy Images; Machine Learning (cs.LG); photon dose; 03 medical and health sciences; Algorithmic autofocusing methods; Space Science; Single-Shot Autofocusing; autofocusing methods; FOS: Electrical engineering, electronic engineering, information engineering; out-of-focus plane; microscopy system; sample areas; imaging times; Image and Video Processing (eess.IV); microscopy image; tissue sections; sample tilt; autofocusing framework; snapshot autofocusing; Electrical Engineering and Systems Science - Image and Video Processing; distance sensors; 004; sample volume; Deep Learning Autofocusing; learning-based offline autofocusing.; Medicine; Biotechnology; Biological Sciences not elsewhere classified; Physics - Optics; Optics (physics.optics)

We demonstrate a deep learning-based offline autofocusing method, termed Deep-R, that is trained to rapidly and blindly autofocus a single-shot microscopy image of a specimen that is acquired at an arbitrary out-of-focus plane. We illustrate the efficacy of Deep-R using various tissue sections that were imaged using fluorescence and brightfield microscopy modalities and demonstrate snapshot autofocusing under different scenarios, such as a uniform axial defocus as well as a sample tilt within the field-of-view. Our results reveal that Deep-R is significantly faster when compared with standard online algorithmic autofocusing methods. This deep learning-based blind autofocusing framework opens up new opportunities for rapid microscopic imaging of large sample areas, also reducing the photon dose on the sample.<br/>27 pages, 8 figures, 9 supplementary figures, 2 supplementary tables<br/>