Multi-view shape constraint system and method for specific coronary artery calcification analysis

Abstract

The invention discloses a multi-view shape constraint system, which comprises a slicing unit for slicing a received image to obtain a coronal slice, a sagittal slice and an axial slice; a first expansion Unet unit, a second expansion Unet unit and a residual expansion network RDN unit for receiving the coronal slice, the sagittal slice and the axial slice respectively; a multi-view shape constraint unit used for integrating different view information through multi-view attention learning to generate multi-view attention features; and a multi-task learning unit that estimates a related calcification index using the output of the multi-view shape constraint unit. The first expansion Unet unit and the second expansion Unet unit model the coronal slice and the sagittal slice and obtain prediction graphs of the coronal view and the sagittal view as supplementary information of calcification lesions, and the residual error expansion network RDN unit models the axial slice and outputs the prediction graph of the axial slice.

Description

technical field [0001] The present invention relates to the field of medical image processing, and more particularly, the present invention relates to a multi-view shape constraint system and method for specific coronary artery calcification analysis. Background technique [0002] Specific coronary artery calcium analysis provides comprehensive judgment of calcification in different coronary vessels. Reveals the site and extent of stenosis in each coronary artery by simultaneously detecting and directly quantifying arterial-specific calcification. Calcium detection can capture the distribution information, geometric and structural characteristics of calcium lesions, and provide high diagnostic value. Quantitative measurement of arterial-specific calcification can provide the extent of calcification in individual arteries, which correlates closely with the severity of stenosis in these vessels. [0003] However, the problem of specific coronary artery calcium analysis aims ...

AI Technical Summary

Problems solved by technology

[0003] However, the problem of specific coronary artery calcium analysis aims to obtain discriminative feature representations for multiple types of high-dimensional tasks

First, there is a dimensional difference between 2D plane detection and 1D linear quantization metrics

This dimensional difference leads to complex relationships among these multi-type coronary artery calcium indicators, increasing the difficulty of multi-type task-dependent learning and commonality mining

Second, there are distributional differences between different calcium fractions

Existing methods cannot learn a joint distribution for multiple regression tasks because they only focus on one score at a time

In addition, existing multi-task learning methods in other fields cannot give satisfactory results due to the multi-type high-dimensional index estimation for specific arterial calcification analysis.

Images