Nuclear positive cell labeling method

Abstract

The invention relates to the technical field of medical image auxiliary recognition, in particular to a nuclear positive cell labeling method which comprises the following steps: S1, obtaining an immunohistochemical image from a tissue slice; s2, the immunohistochemical image is recognized, and a plurality of nuclear positive cells in the immunohistochemical image are obtained; and S3, labeling the nuclear positive cells to generate a labeled image. The method has the beneficial effects that the immunohistochemical image is directly labeled, so that the problem of registration errors caused by registration by introducing a dyed image in the prior art is avoided, and the labeling precision is effectively improved.

Description

technical field [0001] The invention relates to the technical field of medical image aided recognition, in particular to a method for labeling nuclear positive cells. Background technique [0002] Nuclear positive cells refer to tumor cells that show positive results after immunohistochemical reaction in tumor tissue, which can usually be used to express the positive rate and nuclear positive intensity in the tumor tissue, and then realize the disease process of the tumor representation. Immunohistochemical reaction, the full name of immunohistochemical technology, is a research method that uses the basic principle of immunology: antigen-antibody reaction to make the chromogenic agent of the labeled antibody develop color through chemical methods, and then determine the antigen in the tissue cells. . Since the combination between antigen and antibody is highly specific, this technique is widely used in the detection of tumor cells. [0003] In the prior art, there have be...

AI Technical Summary

Problems solved by technology

However, in the actual implementation process, the inventors found that since the annotation of the tumor area in the prior art is further identified through the registration of the HE stained image and the immunohistochemical image, additional information will be introduced in the registration process. The registration error leads to certain difficulties in the image recognition process, which in turn affects the discrimination accuracy.

At the same time, the existing technology often relies on manual image processing when counting nuclear-positive cells, which cannot effectively deal with different types of images

And, lack the analysis method of the nuclear positive intensity of nuclear positive cell in the prior art

Images