Automatic labeling method and system suitable for multiple types of pathological images

Abstract

The invention provides an automatic labeling method and system suitable for multiple types of pathological images. The method comprises the steps: carrying out classification and recognition of an uploaded pathological image through a trained deep learning model, obtaining a classification result of the pathological image, position coordinates of each classification, and the probability corresponding to each classification, automatically labeling the part identified as a tumor region to generate a labeling result, displaying a labeling result of the tumor region in the pathological image, if a pathological expert considers that the labeling result is wrong, performing operations including adding a label, modifying the label and deleting the label and then re-labeling on the labeling result, and giving an auditing result and a diagnosis suggestion by the pathological expert according to the labeling result. According to an automatic pathological image labeling algorithm based on clustering analysis, XML vector diagram labeling of the pathological image is given, a doctor auditing function is designed, the accuracy of automatic labeling is ensured, and the workload of pathological experts is relieved by combining the automatic labeling algorithm with the actual process of a hospital.

Description

technical field [0001] The invention relates to the technical field of medical image diagnosis, in particular to an automatic labeling method and system applicable to multiple types of pathological images. Background technique [0002] In recent years, with the continuous development of artificial intelligence and deep learning technology, computer-aided diagnosis and treatment has become the focus of medical imaging research. The histopathological image-assisted diagnosis method based on deep learning can speed up the diagnostic efficiency of pathologists and alleviate the scarcity of pathologists, long training period, and general overload of pathology departments. Pathological image diagnosis based on deep learning requires the use of a large number of pathological data sets, but the size of full-slice images is huge, and the labeling work is very time-consuming, resulting in fewer pixel-level labeled data sets in the field of pathological images. Most of the existing la...

AI Technical Summary

Problems solved by technology

Chinese patent CN110659692A realizes the automatic labeling of pathological images, but it does not include the doctor's review function, so the accuracy of automatic labeling cannot be guaranteed

Chinese patent CN110826560A provides a pathological image labeling method for esophageal cancer, but this invention only corresponds to a labeling method for one disease

Images