Intelligent Interventional Ultrasound Fusion and Guidance

图像引导手术使用超声成像要求外科医生将术前成像的解剖学和病理学知识与实时超声图像相结合。这需要专业技能和经验。计算机辅助导航系统旨在通过融合术前图像和空间跟踪超声图像来简化信息融合过程。传统的计算方法，如多模态图像配准，在特定的外科应用中面临着重大挑战，而最近的人工智能（AI）基于深度学习的范例需要非常大的介入成像数据训练规模。该项目计划在新的计算框架中应用深度神经网络，开发一个患者特定的手术图像配准系统。该项目将探索一系列最先进的机器学习方法，包括图像合成，基于内容的检索，领域自适应和基于学习的配准技术，以开发用于腹腔镜肝脏切除术的人工智能手术导航系统。该项目的临床目标是大幅提高手术导航的自动化和精确度，使用已经可用或可行的临床数据进行训练和验证。学生应具有计算建模和现代深度学习方法和工具的能力，精通编码和医学图像计算。

Image-guided surgeryusing ultrasound imaging requiresa surgeon to combine anatomical and pathological knowledge from preoperative imaging with real-time ultrasound images. This requires expert skills and experience.Computer-assisted navigation systems aim tostreamline the information fusion processby fusing the preoperative images with spatially-tracked ultrasound images.Traditional computational methods such as multimodal image registration facesignificant challenges incase-specificsurgical applications, whilst the recentartificial intelligence(AI) paradigm based on deep learningrequiresprohibitively large training sizefor interventional imagingdata.This project plans to apply deep neural networkswithin a new computational frameworkto develop a patient-specific surgical image registration system. The project willexplore a range of state-of-the-art machine learning methods including image synthesis,content-basedretrieval, domain adaptationand learning-based registration techniques, to develop anAI-enabledsurgical navigation system for laparoscopic liver resection.The clinical goal of this project is to dramatically improve status quosurgical navigation performance in automation and precision, using clinical dataeither already available or feasible to obtainfor training and validation.The student is expected tohave an aptitude for computationalmodellingand moderndeep learning methods and tools, proficiency in codingand medical image computing.