基于小波变换的磁共振图像中左心室心肌边界检测增强方法

Wavelet-based enhancement for detection of left ventricular myocardial boundaries in magnetic resonance images.

作者信息

Fu J C, Chai J W, Wong S T

机构信息

Automated Measurement and Diagnostic Systems Laboratory, Department of Industrial Engineering, Da-Yeh University, Taiwan, People's Republic of China.

出版信息

Magn Reson Imaging. 2000 Nov;18(9):1135-41. doi: 10.1016/s0730-725x(00)00202-2.

DOI:10.1016/s0730-725x(00)00202-2

PMID:11118768

Abstract

MRI is noninvasive and generates clear images, giving it great potential as a diagnostic instrument. However, current methods of image analysis are too time-consuming for dynamic systems such as the cardiovascular system. Since dynamic imagery generate a huge number of images, a computer aided machine vision diagnostic tool is essential for implementing MRI-based measurement. In this paper, a wavelet-based image technique is applied to enhance left ventricular endocardial and epicardial profiles as the preprocessor for a dynamic programming-based automatic border detection algorithm. Statistical tests are conducted to verify the performance of the enhancement technique by comparing borders manually drawn with 1. borders generated from the enhanced images, and 2. borders generated for the original images.

摘要

磁共振成像（MRI）是非侵入性的，能生成清晰的图像，使其作为一种诊断工具具有巨大潜力。然而，当前的图像分析方法对于诸如心血管系统这样的动态系统来说耗时过长。由于动态成像会生成大量图像，因此计算机辅助机器视觉诊断工具对于实现基于MRI的测量至关重要。在本文中，一种基于小波的图像技术被应用于增强左心室心内膜和心外膜轮廓，作为基于动态规划的自动边界检测算法的预处理程序。通过比较以下两者进行统计测试，以验证增强技术的性能：1. 由增强图像生成的边界与手动绘制的边界；2. 为原始图像生成的边界。