一种新的线边界检测方法

A Novel Line Boundary Detection Method

提出了一种超声图像的多尺度非线性抑噪的自适应线边界检测方法 ,首先对超声斑点图像进行多尺度分解 ,然后对高频部分应用非线性软阈值方法来抑制斑点噪声 ,再利用逆小波变换重建图像 ,最后采用基于“窄条”的线边界检测方法对降噪图像进行处理 ,以不同方向与大小的“窄条”来近似组织边界 ,其中“窄条”的大小由基于区域增长的局部统计特性决定 .

An adaptive line boundary detection technique combining with multiscale nonlinear thresholding for speckle supressing in medical ultrasonic image is presented. An origin image is transformed into a multiscale domain. The components of high frequency are then processed by a nonlinear soft thresholding method. It is reconstructed by inverse wavelet transform. Finally, it is processed by a line boundary detection technique based on sticks. Sticks which are varied in orientation and size are treated as tissue boundary. The size of sticks is decided by local statistics based on region growing technique. The results of the phantom and issue images show its prominent performance on suppressing speckle noise and enhancing issue boundary.