The 30th International Conference on Advanced Laser Technologies B-I-19
ALT'23
Deep-learning 4D live fluorescence microscopy advancing biomedical
applications
Peng Fei1
1- School of Optical and Electronic Information, Huazhong University of Science and Technology
Main author email address: feipeng@hust.edu.cn
Many important biological phenomena, such as heartbeat, blood flow, organelle interactions, etc., often require rapid and high-resolution observations in 4-dimensional time-space. Current fluorescence microscopy imaging techniques are difficult to solve such challenges due to limited spatiotemporal resolution (limited optical throughput), and the measurements are either fast but inaccurate, or accurate but not fast. This presentation will describe how we combined new design in imaging optics with cutting-edge deep learning to increase the throughput of 3D fluorescence microscopy by 2 orders-of-magnitude and achieve ultra-high spatiotemporal resolution observations at both the tissue and single-cell levels. Our deep learning-enhanced light field and light-sheet microscopy techniques realize three-dimensional high-resolution imaging of millisecond-level biological dynamics, such as heartbeat and neural behavior, in freely-moving samples, as well as super-resolution visualization of the 3D interactions between various organelles in a living cell.