基于圆周光强检测的血管OCT图像导管的去除方法

在血管OCT(Optical Coherence Tomography)图像中,OCT图像中导管在不同位置的成像,很大程度上影响了血管内其他目标检测及识别的结果。为了减少图像中导管对上述研究结果的干扰,提出一种快速检测和识别导管区域的自动化方法。该方法先在图像中心处划分检测圆的圆心坐标变化区域,在极坐标下建立圆周光强检测模型。然后,利用每个检测圆上其半圆周的像素均值差构建描述当前检测圆的特征数据。针对导管成像失真问题,采取动态地变换检测圆的圆心坐标和半径的方法。对4组血管OCT图像(约300幅/组)进行实验,实验结果表明,导管检测的准确率达到97.35%,且平均每幅图像的检测时间约为0.7秒。

Removal Method of OCT Image Catheter Based on Detection of Circumferential Intensity

In the OCT image of coronary artery,the catheter imaging at different locations largely impacts the detection results of other objects inner the vessel.To decrease the effects of catheter and obtain an efficient and accurate image information,an automatic method of rapid detection and identification of catheter regions is proposed.Circle center coordinate transforming area is firstly defined and circumferential intensity detection model was built under polar coordinate.Then,the feature data describing the current detection circle are constructed,and the obtained feature data are statistically analyzed.To deal with the distortion problem of catheter imaging,dynamically transforming the center coordinate and radius of the detection circle is proposed.The method is applied to the detection of catheters in 4 groups of OCT images (about 300 per group).The accuracy was 97.35% with the situation of containing a small amount of residual blood,and the processing time was about 0.7 seconds.