Digital holographic microscopy phase noise reduction based on an over-complete chunked discrete cosine transform sparse dictionary

It is well known that a highly coherent laser is used as the illumination source in the digital holographic mi-croscopy (DHM) recording optical path. Lighting sources and optical components may introduce a lot of coher-ent noise, which is marked as random noise, into the system. In addition, the sample characteristics and the experimental environment also introduce various random noises. Random noise severely degrades the quality and accuracy of the DHM imaging. To reduce the effect of random noise, this paper proposes a noise reduction method based on an over-complete chunked discrete cosine transform (DCT) sparse dictionary, also called a com-pressed sensing (CS) sparse representation reconstruction image method. The experimental results show that the reconstructed phase peak's signal-to-noise ratio (PSNR) improves from 53.1230 to 66.4436. The sample feature similarity (FSIM) improved from 0.8171 to 0.9377. The experimental data fully demonstrate that the proposed method can reduce the effects caused by random noise and effectively retain the sample details. Compared to traditional filter methods, it has advantages of noise reduction performance, detail retention performance and processing efficiency. Therefore, the proposed method in this paper will play a more important role in the DHM phase measurement.