Key points are not available for this paper at this time.
The microscopic evaluation of tissues is greatly aided by histological staining, which plays an important role in diagnostic pathology and life science research, but histological staining is often labor-intensive and expensive. Unlike conventional microscopy methods, multiphoton microimaging does not require any special labeling or staining of the sample prior to observation. Histologically stained pathology images can now be acquired by deep learning thanks to advancements in the field of image generation and image transformation. In this paper, we use deep learning network to perform virtual staining on the acquired multiphoton microscopy images, and the results show that the histological stained images generated by virtual staining can show most of the details in the original images, which is close to the original stained images, and it is feasible to perform virtual staining on multiphoton images using deep learning.
Li et al. (Thu,) studied this question.