An intelligent medical image segmentation method based on three-dimensional reconstruction

Abstract

The invention discloses an intelligent segmentation method of medical image organization based on three-dimensional reconstruction, which comprises the following steps: 1) inputting DICOM medical image sequence, preprocessing the image and adjusting the window; 2) performing wavelet transformation on that image to reduce the image data to 1 / 4 of the original image; 3) performing statistic gray histogram to extract initial seed point and growth threshold; 4) carrying out three-dimensional edge detection on that image to obtain an edge contour map; 5) carrying out three-dimensional space regiongrowth by combining gray level and edge information; 6) Calculating the region attributes after segmentation, optimize the seed point and growth threshold. Through statistical histogram, the inventioncan automatically extract initial seed points required for region growth and threshold required for growth criterion, and then carry out three-dimensional region growth combined with edge contour information of an image, and continuously optimize initial seed points and growth threshold through iterative growth, thereby improving segmentation results. At that same time, the invention reduces theimage data to 1 / 4 of the original image through the wavelet change, thereby shortening the operation time.

Description

technical field [0001] The invention relates to the technical field of medical image segmentation, in particular to an intelligent segmentation method of medical image tissue based on three-dimensional reconstruction. Background technique [0002] Medical image segmentation refers to separating regions with different meanings in medical image data according to a specific property, so that the same regions meet regional consistency. Commonly used medical image segmentation methods include: threshold-based segmentation methods, region-based segmentation methods, edge-based segmentation methods, and cluster-based segmentation methods. [0003] The threshold-based segmentation method is to use one or a group of gray thresholds to mark the image data within the threshold range as the region of interest, and mark the image data outside the range as background data, so as to achieve the purpose of image segmentation. The threshold-based segmentation method has the advantages of si...

AI Technical Summary

Problems solved by technology

[0008] 1. Using prior knowledge for segmentation, for complex medical images, any single segmentation algorithm is difficult to obtain satisfactory segmentation results

[0009] 2. Most of the segmentation algorithms are based on two-dimensional image data, without considering the position information of the image in three-dimensional space

[0010] 3. The current automatic segmentation algorithm is difficult to meet the requirements of segmentation accuracy, so the segmentation process usually introduces manual interaction information, and the segmentation efficiency is low

Images