斑点形变资讯斑点位移资讯.PPT

Ultrasonic Strain Image Made by Speckle Tracking of B-mode Image 學生: 呂仁碩 李承諺 葉文俊 指導教授: 李百祺 Introduction Elasticity imaging can provide a significant adjunct to current diagnostic ultrasonic methods Most elasticity imaging is derived by RF data and few by envelope data. RF data elasticity imaging is more sensitive; envelope data is less noisy under larger strain Strain estimation by RF data Strain estimation by envelope data --Speckle tracking of B-mode image 1D-cross correlation of consequent column of pixels (like A-line of RF data)(Ophir) : lateral displacement is hard to detect. 2D-Block matching algorithm (Levinston, 1994; Yeung, 1997), time costing full search and multi-level methods Goal of this project studying the differences of 1D and 2D methods studying the differences of full search and multilevel speckle tracking method searching other possible strain estimator Materials and Methods Target: B-mode images simulated data set breast phantom clinical breast tumor Method: 1D and 2D speckle tracking Result-I Comparison of Elastogram1 (1D and 2 D) Result-I Comparison of correlation coefficients distribution (1D and 2 D) Result-I Comparison of arrow direction distribution (1D and 2 D) Result-I Improved Speckle Tracking  - Multi-level Result-II Comparison of full search and mult-level speckle tracking method Result-II Result-II Result-II Comparison of full search method of sum of difference and correlation coefficient Result-II Result-II Interpolation (1pixel -- 5 pixels) LSQSE(Least Square Strain Estimator) LSQSE(Least Square Strain Estimator) FFT-based Speckle Analysis 報告內容： 原理說明—FFT之特性及優點 文獻參考—Temporal Stretch Method 方法與步驟 結果與討論 原理說明—FFT之特性及優點 FFT為快速演算法 具有 translation invariant特性(Magnitude) 文獻參考—Temporal Stretch FFT-based Speckle Analysis 由頻域對B-mode影像進行分析 Z軸方向的一段組織反射信號 FFT-based Speckle Analysis 信號與系統理論 ： a(z) ? A(ω)a(kz) ? A(ω/k) FFT-based Speckle Analysis 將此方法推廣至二維 ： 方法步驟—形變分析 使用一系列以ph