Perspective correction for interactive projection during surgery

外科医生需要更好的计算机接口，以便在手术过程中与医学图像进行交互。导航通过 一组医学图像并对其进行操作（缩放、平移、滚动等），外科医生目前被限制使用 传统的鼠标和键盘就在无菌手术区域外的控制室中。在每次互动中 对于这些未消毒的器械，外科医生必须（1）“擦洗”进入控制室， 或（2）口头指示技术员操作工作站。在洗手间进进出出 无菌区是耗时的，并且涉及污染风险和向助手发出指示 从外科医生那里移除直接控制。新西兰科技公司（NZTech）一直在开发一种新的医疗 一种名为“无菌环境下与图片和图像通信系统（PACS）的无接触交互”的设备 （TIPSO）。TIPSO采用环境的实时光学3D传感，使外科医生能够 直接从手术室通过手势和动作以完全无菌的方式操作图像。一个 自适应光投射模块使外科医生能够直观地与“虚拟触觉”菜单进行交互， 投影在手术床上。它通过视觉反馈机制增强，使得外科医生接收到 这些操作的直观确认。这项拨款旨在开发一种新的单应性校正 模型，用于以这样的方式预调整投影的图形内容，使得它们看起来不失真， 外科医生的视角该模型将补偿投影上的变形和不规则性 面该项目开发的技术减轻了感染的可能性， 治疗计划制定延迟，从而提高患者安全性并降低医疗保健成本， 相关并发症。该解决方案有助于实现更直观、更高效的图像处理工作流程， 手术环境，同时也成为完善现有人机的重要垫脚石 交互范例构成了当今机器视觉研究的前沿。

Surgeons need better computer interfaces to interact with medical images during surgery. To navigate through a set of medical images and manipulate them (zoom, pan, scroll, etc.), surgeons are currently constrained to use a traditional mouse and keyboard in the control room just outside the sterile surgical field. In each interaction with these unsterile devices, the surgeon has to either (1) "scrub out" to enter the control room and "scrub back in" afterwards or (2) verbally instruct a technologist to operate the workstation. Scrubbing in and out of the sterile field is time-consuming and involves contamination risks and issuing instructions to an assistant removes direct control from the surgeon. NZ Technologies Inc. (NZTech) has been developing a novel medical device, called 'Touchless Interaction with Picture and Archiving Communication Systems (PACS) in Sterile Operations' (TIPSO). TIPSO employs real-time optical 3D sensing of the environment to enable surgeons to manipulate the images by using hand gestures and motions in a fully sterile manner - directly from the OR. An adaptive light projection module enables surgeons to intuitively interact with a 'virtual tactile' menu that is projected over the surgery bed. It is enhanced with a visual feedback mechanism such that the surgeon receives intuitive confirmation of these manipulation actions. This grant aims to develop a new homographic correction model for pre-adjusting the projected graphic contents in such a way that they appears undistorted from the perspective of the surgeon. This model will compensate for deformations and irregularities on the projection surface. The technology developed as a result of this project mitigates the possibility of infection and avoids delays in therapeutic plan formulation, thereby improving patient safety and decreasing healthcare costs and related complications. This solution facilitates a more intuitive and efficient image manipulation workflow in a surgery environment, while also becoming an important stepping stone in refining the existing human-machine interaction paradigms that form the forefront of machine vision research today.