Method and system for automatically analyzing panoramic image of digital pathological section

Abstract

The invention provides a method and a system for automatically analyzing a panoramic digital pathological section image. Color enhancement, normalization and standardization processing are carried out on a pathological image, and an interesting feature extraction region in the image is acquired; the color, the morphology, the texture and the space distribution feature of the pathological section in the interesting feature extraction region in the image standardization module are detected; according to the detection result of the image feature extraction module, automatic classification and quantitative index calculation are carried out on the pathological image; according to the classification result of the automatic classification module, diagnosis by a medical staff is assisted. Thus, automatic analysis on the panoramic image of the digital pathological section can be realized accurately and in real time.

Description

technical field [0001] The invention relates to the technical field of medical image processing, in particular to an automatic analysis method and system for panoramic images of digital pathological slices. Background technique [0002] At present, clinical pathological diagnosis is the most intuitive and accurate gold standard for disease diagnosis. Through the observation of diseased tissue under a microscope, the cancer can be diagnosed and the malignancy of the cancer can be obtained through analysis. Post is expected to provide the best approach. With the rapid development of computer and information technology, high-resolution panoramic imaging of pathological slices has been realized, and digital pathological slice data have been obtained. However, due to the complexity, variability and data volume of pathological images, the existing Pathological image diagnosis still relies on manual image reading. On the one hand, the work is heavy and the professional requirement...

AI Technical Summary

Problems solved by technology

[0005] (1) The complexity of pathological images, due to the influence of image acquisition, such as the integrity of slice making, uniformity of staining, etc., and not only cancer cells in pathological images, but also other tissues such as fat, blood vessels, and fibers interfere with each other, resulting in Image analysis is very difficult

[0006] (2) The variability of pathological images, the shape and distribution characteristics of different cancer cells are inconsistent, and the shape of cancer cells is variable and irregular, the performance of cells in different division stages is variable, and the overlapping of nuclei and other factors affect the pathological image. Analysis poses a big challenge

[0007] (3) Computational complexity. Since the single image data of panoramic pathological images is very large, tens of millions to hundreds of millions of pixels, how to achieve fast and real-time calculation is a technical difficulty of pathological image analysis algorithms

[0008] These technical problems restrict the actual clinical application. According to the published literature, there is still no systematic solution to realize the automatic analysis of panoramic pathological slice images and extract quantitative analysis indicators, especially difficult to solve the pathological feature learning algorithm. The contradiction between real-time and detection accuracy in

Images