Enhancing multimodal binary mask modulation imaging efficiency with momentum acceleration

Multi-amplitude modulation imaging algorithms begin with an initial set of guess functions. The detected intensity information is iteratively fed into a correction loop, refining the guess functions until convergence is achieved. To enhance imaging speed, we propose a high-performance method that employs a three-frame modulation mask and incorporates momentum-based acceleration for faster encoding. The incorporation of momentum effectively introduces an inertial property to the iterative reconstruction process, addressing the issue of reduced reconstruction speed caused by diminished prior modulation information. Compared to conventional approaches, the proposed method demonstrates superior performance in both data acquisition time and algorithmic reconstruction efficiency. Furthermore, by leveraging amplitude modulation devices, this technique may enable breakthroughs in high-speed complex amplitude imaging for short-wave applications, with numerical simulations confirming its feasibility.