Phase retrieval with dynamic linear combination in multiple intensity measurements

In the present methods of multiple intensity measurements, the recorded different intensity patterns are supposed to have the identical quality and the contribution of different images to the result is the same. However, intensity patterns are of different quality, since the raw patterns are affected by environment noise and focusing distance in the process of data acquisition. The equal weight scheme in data processing is not optimum in the multiple intensity measurements. A novel, to the best of our knowledge, phase retrieval algorithm with dynamic linear combination is proposed to improve resolution and robustness. In the algorithm, each pattern is assigned different weight according to its quality during reconstruction. The image quality evaluation function is applied to determine the weight coefficient. A higher-resolution image is reconstructed by the linear combination of all images, which is superior to the conventional algorithm.