Research on multi-point light focusing based on complex neural network

The research of light focusing and imaging through scattering media is currently a popular topic. Many focusing technologies have been developed, such as transmission matrix method, phase conjugation method, iterative optimization method, etc. However, these methods have some limitations. At present, deep learning methods have been widely used in the field of image restoration, and have strong inverse restoration capabilities. Yet, the application of this method in the field of optical focusing is still relatively limited, and the performance is not ideal. In this letter, a method based on complex neural network is proposed, and the multi-point focusing of light passing through the scattering medium is numerically simulated. Since the complex information of the network is not reduced, compared with the real number neural network method, this method more accurately simulates the recovery process of light scattering, and can obtain multiple light focal points with high enhancement at the same time.