Construction of detection and classification model of pathological squamous epithelial cells

Abstract

The invention discloses construction of a detection classification model of pathological squamous epithelial cells. The construction comprises construction of an abnormal cell detection model, construction of a view diagram judgment model and construction of a sample judgment model. The three models in the detection and classification model are linked with one another and progressive layer by layer, the result is subjected to optimization detection, detail re-diagnosis, integration re-diagnosis and the like on the detection basis of the previous model, multiple control over the diagnosis result on the whole is achieved, the accuracy of the diagnosis result is guaranteed, and a perfect sample diagnosis method is obtained.

Description

technical field [0001] The invention relates to the field of detection and classification of cells, in particular to the construction of a detection and classification model of pathological squamous epithelial cells. Background technique [0002] Squamous epithelial cells, also known as squamous epithelial cells, mainly come from the lower part of the ureter, bladder, urethra, vaginal surface, and uterus. The surface is covered with squamous epithelial cells. The growth and differentiation of this epithelium are mainly affected by estrogen produced by the ovary. The role of progesterone is to promote the shedding of epithelial cells. Squamous cell lesion is a common change in the process of cervical liquid-based cytology, and it is a normal change of cervical cells. Only when there is atypical squamous cell lesion can the possibility of precancerous lesions be considered. Therefore, for pathological squamous epithelial cells It is difficult to detect and classify, and manua...

AI Technical Summary

Problems solved by technology

Squamous cell lesion is a common change in the process of cervical liquid-based cytology examination. It is a normal change of cervical cells. Only when there is atypical squamous cell lesion can the possibility of precancerous lesions be considered. Therefore, for pathological squamous cell lesion It is difficult to detect and classify, and manual diagnosis requires rich experience, but it still takes a lot of time, and there will be misjudgments, resulting in the inability to accurately screen

[0003] At present, the use of computer-aided detection of squamous epithelial cells based on image features can improve the diagnostic efficiency, but due to the unreasonable model construction, the diagnostic results cannot guarantee high accuracy, and there is a lack of diagnostic methods for the entire sample

Images