Biopsy Needle Dynamic Puncture Algorithm Guided By Ultrasound Images

With the development of science and technology,puncture interventional surgery has been widely used in human biopsy,lumbar puncture,and target drug injection.In the guidance of interventional surgery,ultrasound images are more and more widely used due to the advantages of non-invasive invasion,fast imaging speed,and high cost performance.However,when performing biopsy surgery,du e to the influence of natural breathing movement,the position of the target and the obstacle will also change accordingly.Therefore,how to realize the preoperative puncture path planning of the puncture probe and the fine adjustment of the intraoperative puncture path are increasingly The more important.In response to the above problems,this paper combines the motion information of obstacles and targets with the puncture path planning of the puncture probe,and proposes a method of dynamic puncture path planning based on ultrasound image-guided puncture probe.which the target detection,automatic target segmentation,target tracking and flexible needle puncture trajectory planning based on target and obstacle movement information are deeply studied.First,the blood vessel area is automatically extracted based on the ultrasound image.For the collected ultrasound images,morphological filtering is used to remove the annotation information in the ultrasound images;Apply the existing deep learning-based target detection algorithm to the extraction of the region of interest(ROI)in the ultrasound image and improve it.The improved target detection algorithm is used to extract the blood vessel area in the liver ultrasound image and the extracted blood vessel area.The extracted ultrasound image of the blood vessel area will be directly used for subsequent blood vessel segmentation.Secondly,Perform the blood vessels segmentation based on the extracted ultrasound images.According to the characteristics of the ultrasound image,the histogram equalization is used to preprocess the ultrasound image;the ultrasound image segmentation algorithm based on U-net is used to segment the blood vessel from the extracted ultrasound image of the blood vessel area,so as to realize the automatic recognition of the blood vessel.The size and position information of the identified blood vessels will be used as obstacles to be avoided in path planning and used to guide subsequent path planning.Again,the blood vessels in the ultrasound image are tracked.By analyzing the basic characteristics of the long-term and short-term memory network(LSTM)and applying it to the prediction of target information in ultrasound images,a LSTM-based blood vessel information prediction algorithm is proposed;based on the existing blood vessel information in the ultrasound image,the follow-up is predicted Blood vessel information in ultrasound images.This information will be used as obstacle movements to be considered in path planning and to guide subsequent puncture path replanning.Finally,based on the identified obstacle information and the predicted obstacle motion information,a real-time puncture path planning algorithm based on the current state is proposed.By analyzing the basic characteristics of the rapid search random tree(RRT)algorithm,the RRT algorithm is improved,and the reachable guidance and greedy heuristic strategies are introduced;according to the real-time position of the identified targets and obstacles,combined wi th the old point tracking strategy(OPTS)and extreme The trend expansion strategy(ETES)uses the reachable guidance and greedy-inspired fast search random tree(GHRG-RRTs)algorithm to plan the needle path in real time,and proposes a suitable cost funct ion according to the actual puncture characteristics of the puncture probe to determine the optimal Puncture path.