Medical image multi-threshold segmentation method based on interval iteration

Abstract

The invention provides a medical image multi-threshold segmentation method based on interval iteration. The method comprises the following steps: 1, obtaining a base layer of a source image by adopting a mixed L1-L0 layer decomposition method; secondly, an Otsu multi-threshold segmentation method based on interval iteration is designed to segment a source image and a base layer of the source image, and two different segmentation results are obtained; and thirdly, constructing a segmentation fusion method to merge different segmentation results to obtain a final image segmentation result. According to the method, the optimal threshold value is obtained through interval iteration of the gray level histogram, and mistakenly divided pixels are re-divided by adopting the fusion strategy, so that the accuracy of image segmentation can be improved. Experimental results show that the medical image multi-threshold segmentation method based on interval iteration can process the medical image and has robustness to noise.

Description

technical field [0001] The invention belongs to the field of medical image segmentation, in particular, an Otsu multi-threshold segmentation method based on interval iteration is designed to realize brain MR image segmentation, and a fusion strategy is used to improve the accuracy of image segmentation. Background technique [0002] Image segmentation technology is a key step from image processing to image analysis and plays an important role in image recognition and computer vision. Image segmentation refers to dividing the image into several disjoint areas according to some inherent characteristics of pixels in the image (such as grayscale, color, spatial texture and geometric shape, etc.), so that the characteristics of the same area are consistent or similarities, while the features between different regions have obvious differences. The purpose of image segmentation is to classify each pixel in the image to extract regions of interest (ROI). [0003] In the field of m...

AI Technical Summary

Problems solved by technology

[0005] However, the operating efficiency of the traditional multi-threshold segmentation method is low, and its time complexity will increase exponentially with the increase in the number of thresholds.

In addition, medical images are complex, with uneven gray levels, blurred boundaries between different tissue regions, complex shapes and structures, and artifacts, which bring great difficulties to the accurate segmentation of images.

Therefore, multi-threshold segmentation of medical images remains a challenging task

Images