Fig. 1. Schematic of high-speed optofluidic imaging

Fig. 2. Fluorescence imaging technologies. (a) A multi-focal confocal microscope based on structured light for high-resolution optofluidic imaging of living cells^[20]; (b) schematic of microfluidic imaging based on fluorescence lifetime microscopy^[22]; (c) schematic of optofluidic imaging based on light-sheet fluorescence microscopy^[25]; (d) optofluidic images of a single K562 cell with green and blue fluorescent staining obtained by light-sheet fluorescence microscopy, scale bar is 5 μm^[25]

Fig. 3. Bright-field imaging technologies. (a) Schematic of optical time-stretch (OTS) imaging^[27]; (b) bright-field images of single platelets, aggregated platelets, and white blood cells obtained by high-speed optofluidic OTS imaging^[29]; (c) schematic of high-speed bright-field imaging based on frequency-division multiplexing (FDM)^[14]; (d) bright-field images of single platelets and platelet aggregates obtained by high-speed optofluidic FDM imaging^[14]

Fig. 4. Quantitative phase imaging (QPI) technology. (a) Schematic of QPI based on OTS^[35]. (b) typical phase images of single platelets and platelet aggregates^[35]

Fig. 5. Nonlinear optical imaging technologies. (a) Schematic of optofluidic imaging based on stimulated Raman scattering (SRS)^[36]; (b) SRS optofluidic images of whole blood cells, PBMCs, Jurkat cells, and HT29 cells^[36]; (c) schematic of high-speed optofluidic imaging based on four-wave mixing (FWM) and second harmonic generation (SHG)^[39]; (d) FWM (left), SHG (middle), and merged (right) images of C. Zofingiensis cells^[39]

Fig. 6. Microfluidic focusing technologies commonly used in high-speed optofluidic imaging. (a) Hydrodynamic focusing; (b) dielectrophoresis (DEP)-based focusing; (c) acoustic focusing

Fig. 7. Applications of deep learning in high-speed optofluidic imaging. (a) Images deblurring of K562 cell using convolutional neural network (CNN)^[63]; (b) real-time single cell image acquisition and classification using CNN^[64]; (c) classification of single platelet, platelet aggregates, and white blood cells using CNN^[31]; (d) generation of high-resolution images of Jurkat cells from the low-resolution images using generative adversarial network (GAN)^[17]