3D Gabor wavelet based vessel filtering of photoacoustic images.
3D Gabor wavelet based vessel filtering of photoacoustic images.
复制标题
基于 3D Gabor 小波的光声图像血管滤波。
DOI：
10.1109/embc.2016.7591576
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Saijo Y
中科院分区：
文献类型：
--
作者：
Ul Haq I;Nagoaka R;Makino T;Tabata T;Saijo Y
血管图像的滤波和分割是医学成像中的一个重要问题。在许多医学应用中，血管的可视化对于早期诊断和治疗至关重要。研究了在三维光声显微镜（OR-PAM）中，利用Gabor小波增强血管效应，同时消除由于探测器尺寸、灵敏度和孔径等因素引起的噪声。由于血管通常在一定半径范围内变化，因此对小波滤波和基于Hessian的方法进行了详细的多尺度分析，以提取不同尺寸的血管。该算法首先增强图像中的血管，然后通过在图像中的每个体素的局部Hessian矩阵的特征值分解来分类管状结构。该算法在无创实验上进行了测试，结果表明，该算法在增强光声图像中的血管方面取得了明显的效果。
Filtering and segmentation of vasculature is an important issue in medical imaging. The visualization of vasculature is crucial for the early diagnosis and therapy in numerous medical applications. This paper investigates the use of Gabor wavelet to enhance the effect of vasculature while eliminating the noise due to size, sensitivity and aperture of the detector in 3D Optical Resolution Photoacoustic Microscopy (OR-PAM). A detailed multi-scale analysis of wavelet filtering and Hessian based method is analyzed for extracting vessels of different sizes since the blood vessels usually vary with in a range of radii. The proposed algorithm first enhances the vasculature in the image and then tubular structures are classified by eigenvalue decomposition of the local Hessian matrix at each voxel in the image. The algorithm is tested on non-invasive experiments, which shows appreciable results to enhance vasculature in photo-acoustic images.