Momentum Acceleration Fourier Ptychographic Microscopy

Fourier Ptychographic Microscopy(FPM) is a super resolution imaging method, which recovers a high resolution object image from a series of original low resolution images, breaking the mutual restriction between high resolution and large field of vision. However, the actual reconstruction process is easily affected by noise and involves processing a large amount of data. The system needs a lot of iteration time and may fall into local minimum value. In this paper, an improved algorithm of momentum accelerated Fourier ptychographic microscopy is proposed. Based on the adaptive step-size algorithm, the momentum used to accelerate the weight training in the neural network is introduced, the changes of the previous objective function are added to the current iteration to improve the convergence speed and reconstruction quality of the algorithm reconstruction. Simulation results show that the proposed method is effective.