血管网光声成像的机器学习抗散射仿真研究  

Anti-scattering simulation of photoacoustic imaging of vascular network based on machine learning

作　　者：廖江南 高雅[1] 徐文逸 解维娅 程茜[1] LIAO Jiangnan;GAO Ya;XU Wenyi;XIE Weiya;CHENG Qian(Institute of Acoustics,School of Physical Science and Engineering,Tongji University,Shanghai 200092,China)

机构地区：[1]同济大学物理科学与工程学院声学研究所,上海200092

出　　处：《声学技术》2023年第6期757-763,共7页Technical Acoustics

基　　金：国家自然科学基金(12034015)。

摘　　要：乳腺癌已成为全球女性发病率最高的肿瘤疾病,微血管成像对乳腺癌的治疗方案和预后有重要意义。光声层析成像术(Photoacoustic Tomography,PAT)可有效对乳腺癌内微血管网进行成像,但肿瘤组织内部的异质微结构和钙化点的散射对成像质量影响较大。针对该问题,文章基于U-Net的卷积神经网络对不同颗粒散射条件下软组织中血管网图像散斑开展仿真研究。仿真结果表明,该神经网络可以学习光声散斑图像和成像目标之间的映射关系,提取出隐藏在噪声中的血管光声信号,并重建出轮廓清晰、背景清晰的高质量血管图像,表明U-Net网络可以从高度模糊的散射图像中提取出有效的光声信息,实现目标图像的高清重建,在乳腺癌的诊断成像中具有广阔的应用前景。Breast cancer has become the most prevalent neoplastic disease in women worldwide,and microvascular imaging is of great significance for treatment and prognosis of breast cancer.Photoacoustic Tomography(PAT)can effectively image the microvascular network in the breast,but the scattering from heterogeneous microstructures and calcified spots within the tumor tissue has a great impact on the imaging quality.To address this issue,a simulation study is conducted on the scattering of vascular network images in soft tissues under different particle scattering conditions based on the convolutional neural network of U-Net.The simulation results show that the neural network can learn the mapping relationship between the photoacoustic speckle images and the imaging target,extract the photoacoustic signal of the blood vessel hidden in the noise,and reconstruct the high-quality blood vessel image with clear contour and background,which has a broad application prospect in breast cancer diagnostic imaging.

关 键 词：光声成像 机器学习 血管网成像 

分 类 号：O426.1[理学—声学]