Multi-dimensional Multi-parametrical Functional Reconstruction Of Ultrasound Heart Images

The accurate diagnosis and efficient treatment of cardiopathy is of great social significance. But traditional three-dimensional ultrasound reconstruction system can just depict three-dimensional anatomical structure,so it's very difficult to give an accurate assessment of the functionality of the heart. In this article,based on the accurate reconstruction of three-dimensional anatomical structure,the technique of 3D reconstruction and Doppler ultrasound imaging is combined to implement multi-dimensional multi-parametrical functional reconstruction of ultrasound heart images.In order to perfume functional reconstruction,a 3D reconstruction of anatomical structure from ultrasound medical images is implemented at first. Multiplane transesophageal rotational scanning method is used to acquire original pulse-echo B-scan ultrasound images while echocardiogram is recorded synchronously. According to the characteristic of speckle noise in ultrasound images,adaptive weighted median filter is used to suppress speckle in these images. Then the images at the same time phase in echocardiogram are reorganized according to their corresponsive acquisition angle after the extraction of region of interest. A direct matching interpolation method is utilized to get regular volume data from discrete distribution of ultrasound information points. This volume data can be visualized with surface rendering and volume rendering algorithm. Thus,the reconstruction of the anatomy in a certain time phase at echocardiogram is implemented. If all the rendered results are linked and displayed dynamically according to the time sequence at ECG,dynamic reconstruction of anatomy is accomplished.Then,the grayscale anatomical structure and color-coded functional parameter were seperated from Doppler ultrasonic medical images based on superposition principle and colorbar analysis. The method of artifical neuralnetwork combined with look up table was developed to map color code information to its original functional parameter. This method can acquire both mapping precison and mapping velocity,so it's quite suitable for functional reconstruction of Doppler ultrasonic medical images.At last,multi-dimensional multi-parametrical functional reconstruction was created by the combination the techniques of traditional three-dimensional reconstruction and Doppler medical imaging. The key problem of vector interpolation and fusion imaging in this procedure is solved. Thus,three-dimensional distribution of functional parameters,such as velocity,acceleration or energy of tissue movement and velocity of blood flow,were reconstructed,and the dynamic fusion visualization of the reconstructed functional parameter with its corresponding 3D anatomical structures was realized.Clinical human experiments showed that in vivo functional parameters could be shown in 3D space simultaneously with anatomical structure of the heart,and the relationship between functional structure and functional parameters was consistent with cardiology. The images rendered in the present thesis could show more medical information compared with that in traditional methods.