Digital pathological image H&E dyeing restoration method based on strong reversible adversarial network

Abstract

The invention relates to a digital pathological image H&E dyeing restoration method based on a strong reversible adversarial network. The method comprises the following steps: training a strong reversible adversarial network comprising two generators and two discriminators; and performing H&E dyeing restoration on the digital pathological image through the trained strong reversible adversarial network. The training process of the strong reversible adversarial network comprises: making image data sets before and after fading, then continuously training a forward process and a reverse process, calculating a total loss function until the total loss function and the quality of the image generated by the generator reach a preset reconstruction effect, stopping training, and obtaining optimal convolution weight parameters and offset parameters of a generator and a discriminator. The method is beneficial to color information restoration of the faded digital pathological image.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a method for restoring H&E staining of digital pathological images based on a strong reversible confrontation network. Background technique [0002] Hematoxylin-eosin staining (hematoxylin-eosin staining), referred to as H&E staining, is the most common staining method in paraffin section technology. The hematoxylin staining solution is alkaline, which mainly makes the chromatin in the nucleus and the nucleic acid in the cytoplasm stain purple blue; eosin is an acid dye, which mainly makes the components in the cytoplasm and extracellular matrix red. H&E staining is the most basic and widely used technical method in histology and embryology research. The H&E stained glass slides are scanned by a digital scanner to form digital pathological images that can be displayed by a computer. However, prepared sections tend to fade over time. [0003] In the prior a...

AI Technical Summary

Problems solved by technology

Although the third method has a certain improvement in processing speed and color quality, the result is still unsatisfactory, and it does not solve the problem of dyeing normalization.

Images