Image segmentation method and device and storage medium

Abstract

The invention discloses an image segmentation method and device and a storage medium, and belongs to the technical field of image processing. The method comprises the steps of obtaining a dimension reduction image according to a foreground image; determining a plurality of first segmentation thresholds according to the dimension reduction image; determining a plurality of first segmentation regions according to the plurality of first segmentation thresholds and the foreground image, the inter-class variance of the plurality of first segmentation regions being the maximum and the sum of intra-class variances being the minimum; and determining a first target area and a second target area according to the plurality of first segmentation areas. According to the invention, the dimension reduction image is processed; a plurality of first segmentation thresholds are determined adaptively, furthermore, segmentation is carried out based on the first segmentation threshold value, so that the between-class variance of each segmented region is maximum, the sum of intra-class variances is minimum, and the pixel points with relevance can be segmented into one region as much as possible when thecondition is met, and thus the obtained first target region and the obtained second target region are more accurate.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image segmentation method, device and storage medium. Background technique [0002] Breast mammography is an important means of breast cancer detection. According to the shooting angle, mammography images can be divided into different bitmaps such as CC (Craniocaudal, head and tail), MLO (Mediolateral Oblique, inner and outer oblique). In addition to the breast area, the MLO bitmap usually includes a part of the pectoral muscle area, and the pectoral muscle area will interfere with the detection of breast lesion tissue and the density calculation of breast lesion tissue, which will further affect the detection results of breast cancer. Therefore, it is necessary to Segment the breast area and pectoralis area in the MLO bitmap. [0003] Taking the mammography image as the image to be segmented, the breast area as the first target area, and the pectoralis area...

AI Technical Summary

Problems solved by technology

[0007] Due to the limitations of the Otsu algorithm itself, each segmented area obtained by using the Otsu algorithm can only meet the requirements of the segmentation threshold, resulting in inaccurate segmentation of the first target area and the second target area

Images