Figure | Framework and performance of the MPPN-RW method for optical conveyor belt generation. (IMAGE)
Caption
Figure | Framework and performance of the MPPN-RW method for optical conveyor belt generation. a, Schematic illustration of the MPPN-RW framework. The target optical field is defined by a desired trajectory, where the corresponding phase is generated via a trajectory-informed phase synthesis strategy. The intensity and phase components are encoded and fed into the MPPN-RW network. By incorporating multiple priors, including physical propagation, phase periodicity, and smoothness constraints, the network optimizes the holographic phase. The reconstructed light field is obtained through a physical forward model, and the loss is iteratively minimized to achieve high-fidelity optical field reconstruction. b, Experimental demonstration of optical conveyor belts for particle transport. Different types of particles, including polystyrene (PS), gold (Au), silica (SiO2), and yeast cells, are transported along a flower-shaped trajectory. The left column shows the reconstructed light field, while the right columns present time-lapse snapshots of particle motion along the predefined path, demonstrating stable and continuous transport. c, Quantitative performance comparison with existing methods. The proposed MPPN-RW method achieves significantly improved uniformity in intensity distribution compared with DeepCGH, RP-DIFT, and RPESO-GPOV. Statistical results show enhanced performance across different metrics, indicating superior reconstruction fidelity and transport stability.
Credit
Yanan Cai et al.
Usage Restrictions
Credit must be given to the creator.
License
CC BY