The Study On B-ultrasound Image Registration Of HIFU Monitoring Based On Ultrasonic Speckle

During clinical therapy of Ultrasound-Guide HIFU system, doctors need compare B-mode ultrasound images of different positions to locate the treatment target. As the presence of phantoms and other ultrasound noises, these will increase the difficult of B-mode ultrasound image registration when probe moves form one place to another. In order to assist clinical doctors locate the treatment target more exactly by contrast different B-mode ultrasound images, this study designed a novel B-mode ultrasound image registration based on ultrasonic speckle. According to the pseudo-random feature of speckle in envelope-detected image the regions of speckle could be the feature points of B-mode ultrasound image.Defining the gray range of speckle by statistical method and constructing speckle binary image through extracting pixels in the gray scope the B-mode ultrasound images registration can be achieved by register speckle binary images. When the distance between B-mode ultrasound probe and bovine liver in vitro varyed, we collected six Groups of B-mode ultrasound images and used the image registration based on ultrasonic speckle.Through analyzing characteristics of B-mode images based on the principles of ultrasound speckle,we used the fourier transform phase correlation method and speckle feature points,at last we could registrate B-mode image accurately. The experimental results show that this method has higher robustness and accuracy,and also proved the speckle can be used as feature points of B-mode ultrasound image.The fisrst innovative is that we use ultrasound speckle and speckle theory of pseudo-random characteristics to extract the B mode images scattered spots as feature points for B-image registration of HIFU monitoring;The second innovative is that we matching Ultrasound speckle feature point and registration based on Fourier phase correlation to registrate B mode images;The third innovative is that we can registrate the B mode images even if there is artifacts and noise.