Optical differentiation based on weak measurements
An Wang, Junfan Zhu, Lan E Luo, Xiong Liu, Ling Ye, Zhiyou Zhang, Jinglei Du · Optics Letters · 2022
Optical differentiation shows much potential to be applied in computation due to its strong parallelizability. Currently, each optical differential method can only obtain partial differential information. Here, we propose a general approach to obtain complete differentiation. Compared to previous methods, we can separately obtain the differentiation of amplitude and phase, reserve the negative value of the differentiation, and acquire the differentiation in arbitrary directions of the two-dimensional field. We measure the differentiation of the Gaussian beam to demonstrate this method. A practical experiment of identifying the move direction of the motion-blurred image is also presented to verify the practicability of our method. Our method can further be applied to intelligence algorithms, image identification, and optical analog computing.