Extraction method for region of interest of breast three-dimension image

Abstract

The invention discloses an extraction method for a region of interest (ROI) of a breast three-dimension image. The method includes: performing denoising processing on the breast three-dimension image, dividing the breast three-dimension image into k cross sections in a Z-axis direction and selecting a two-dimension image I<xy> on the cross sections; through the grey information of the two-dimension image, respectively determining internal boundaries and external boundaries of the ROI in an X-axis direction and external boundaries of the ROI in a Y-axis direction; and selecting a two-dimension image I<yz> on a breast sagittal plane and respectively determining boundaries of the ROI in the Z-axis direction through the two-dimension image I<yz>. The extraction method for the ROI of the breast three-dimension image combines with the grey characteristic of the breast three-dimension image so that full-automatic extraction of the ROI is realized accurately and effectively.

Description

Technical field [0001] The present invention relates to the field of image processing, in particular to a method for extracting a region of interest in a three-dimensional breast image. Background technique [0002] Breast cancer is the most common disease among middle-aged and elderly women, and the morbidity and mortality are increasing year by year. Early detection and diagnosis of breast cancer can effectively improve the cure rate of breast cancer. In clinical practice, magnetic resonance images have been widely used for breast cancer risk prediction and early diagnosis. The computer-aided diagnosis system can help radiologists improve the accuracy and efficiency of three-dimensional magnetic resonance images. In a computer-aided diagnosis system based on three-dimensional magnetic resonance images, the automatic segmentation and registration of breasts are the two most basic tasks, and the automatic extraction of regions of interest (ROI) including breasts can greatly impr...

AI Technical Summary

Problems solved by technology

The automatic segmentation method based on grayscale or gradient is not accurate due to the diversity of grayscale or gradient and possible low contrast.

However, the model-based segmentation method needs a large number of training samples to obtain better results, and the acquisition of these training samples is a more difficult and complicated task.

Images