| In order to fully combine the advantages of Coronary angiography(CAG)and Intravascular Ultrasound(IVUS)images produced by the two imaging techniques commonly used in the diagnosis of cardiovascular diseases,which means,the direction and shape of blood vessels in the heart can be observed clearly by CAG images,and the corresponding IVUS sequence frame can also be observed,the combination of the two kinds of image is needed.At present,when studying the fusion of CAG image and IVUS image at home and abroad,we choose the ideal situation,and use the center line of blood vessel in CAG image for three-dimensional reconstruction.In the actual operation,the vessel center skeleton is extracted or the curve is fitted to replace the vessel center line,but the trajectory of the transducer on the IVUS guide wire cannot be guaranteed to coincide with the vessel center line in the process of retracting,the centerline of the vessel does not represent the retracting trajectory of the transducer on the IVUS guide wire.In order to carry out accurate co-registration,this paper restores the trajectory of the ultrasonic transducer on the IVUS guidewire to the greatest extent,that is,uses an interactive segmentation method to extract the IVUS guide wire trajectory in the CAG image.In the implementation of interactive segmentation,the original intelligent scissors algorithm is improved while a new feature cost function is constructed by using 5 different image features.The experiment proved that the segmentation operation is faster and the segmentation effect is better after the improvement.The co-registration can be divided into three steps:1.Interactive segmentation of the guidewire:Use the improved intelligent scissors algorithm to interactively extract the IVUS guidewire traJ ectory in the CAG images of the same vessel under two different projection angles,respectively;2.Three-dimensional reconstruction of the guidewire:use the two trajectories to perform three-dimensional reconstruction of the ultrasonic transducer retracement trajectory.The least-square solution of the 3-d reconstruction system is obtained according to the system parameters and two sets of 2-d coordinates,that is,a set of three-dimensional coordinates based on a world coordinate system.Using the Levenberg-Marquardt Algorithm to optimize the geometric transformation matrix;3.Co-registration of CAG image and IVUS image:The obtained three-dimensional scattered point coordinates are fitted with a b-spline curve based on the approximation of the least square curve,and the best three-dimensional fitting curve is found by the iterative method.According to the property that the transducer retracts along the IVUS guide wire uniformly,the corresponding IVUS frames at each point on the guidewire can be obtained,and then the co-registration can be realized.t has been proved by relevant experiments that in the case of selecting the same cardiac moment with the aid of Electrocardiogram(ECG)gating,the co-registration system proposed in this paper can precisely realize the co-registration among the CAG image,the IVUS image and the vertical section of IVUS.By dragging the sequence of vertical section of IVUS or sliding the sequence of IVUS,the movement of the corresponding position in the actual CAG vessels can be observed,and vice versa,and the whole interactive operation is convenient and quick,the effect of linkage analysis is good.The precise co-registration system based on CAG image and IVUS image can quickly and accurately localize and evaluate the condition of vascular disease.It is of great significance to improve the efficiency of cardiovascular clinical diagnosis and improve the survival rate of cardiovascular patients.At the same time,it can promote the development of precision medicine in the cardiovascular field. |
