实验室显微智能成像系统的研究发表在测量领域老牌期刊Measurement Science and Technology上
Image credit: Results
Digital holographic microscopy (DHM) as a quantitative phase imaging technology, has been widely used in different applications. The phase aberration compensation in off-axis DHM is vital to reconstructing a topographic image with high precision, especially for microstructures with a small background or dense phase distribution. We propose a numerical method based on deep learning in DHM. First, a convolutional neural network (CNN) recognizes and segments the sample and background area of the hologram. Then, Zernike polynomial fitting is executed on the extracted background area. Finally, the whole process of phase aberration compensation is automatically achieved. To obtain a robust and accurate deep learning model for hologram segmentation, we collected many holograms corresponding to several samples that have different morphological characteristics. The experimental results verify that the trained CNN could accurately segment the sample from the background area of the hologram, and this method is applicable and effective in off-axis DHM.
