Thyroid cytology multi-type cell detection method based on deep learning

Abstract

The invention discloses a thyroid cytology multi-type cell detection method based on deep learning, and the method comprises the steps: obtaining a cell pathology slide image of a thyroid cell, obtaining a plurality of target images from the cell pathology slide image through sliding window scanning with selectable step length, the adjacent target images have an overlapping area, the target images are processed to obtain at least one feature image and confidence images, and the number of the confidence images is the same as the number of lesion types of lesion cells of the thyroid; and processing the at least one confidence coefficient image to obtain a plurality of confidence coefficients corresponding to different lesion types, obtaining a classification result matched with the target image based on the plurality of confidence coefficients, and processing the at least one confidence coefficient image based on the classification result to obtain an area matched with the classification result in the target image. Therefore, the detection efficiency of the detection system can be improved, and the marking cost can be reduced when the detection system is trained.

Description

technical field [0001] The present disclosure specifically relates to a method for detecting multiple types of cells in thyroid cytology based on deep learning. Background technique [0002] Liquid-based cytology is a branch of cytopathology, which collects cell samples into liquid fixative, and after staining, the cell samples can be used for observation and diagnosis. Common applications include cervical biopsies, which are used to screen for precancerous cervical lesions that may lead to cervical cancer. [0003] At present, the diagnosis process of the relatively extensive cytopathology slide image is to block the cytopathology slide image through the algorithm model first, and detect the diseased cells in the image block to obtain the location and lesion type of the diseased cells (classification) , and then integrate the classification results. Currently, target detection algorithms are often used to locate and classify diseased cells. However, this method requires ...

AI Technical Summary

Problems solved by technology

[0004] However, the methods disclosed above are only aimed at the probability that an image block belongs to a single lesion, and are not suitable for the classification of multiple lesion types

For some cytopathological slide images (such as cytopathological slide images of the thyroid), one image block usually contains multiple cell clusters, and it is difficult for the cell clusters to completely fall into a single image block, and the classification results of the image blocks are integrated It is easy to cause the target frame to deviate from the actual cell

At the same time, when the cytopathological slide image is divided into non-overlapping image blocks, cell clusters are easily cut by two image blocks, resulting in feature loss and missed detection.

Images