High Precision Surgical Navigation And Its Application

With the rapid development of medical science, surgery put on more and more attention on minimally invasive and refinement. Surgical Navigation System(SNS) matches the patient preoperative image data and intraoperative patient anatomy structure accurately, and plays real-time navigation, which will attribute to higher operation accuracy, smaller operation trauma and less recovery time. According to the needs of SNS, this paper design a new generation of infrared surgery navigation clinical prototype system.This paper studies key technologies of SNS, including binocular camera calibration, stereo calibration, 3-D reconstruction of medical images. In addition, this article does some researches to increase the running speed and the positioning accuracy which are as follows:1. Propose the prediction segmentation algorithm. The surgery instruments'recognition typically uses matching algorithm, which is slowly. Therefore, the prediction segmentation algorithm is proposed, which regarded the trajectory of marks as quadratic curve. It uses the quadratic curve fitted by previous frame to figure out the current prediction coordinate, and then the nearest neighbor algorithm is used to classify the current marks. The algorithm can effectively improve the recognition speed, and can gain high accuracy with template matching algorithm used in combination.2. Propose the high precision reconstruction algorithm of surgery instruments. Traditionally, we recognize the surgery instruments by the marks on themselves, whose coordinates are used to calculate the posture and work point of the instruments. This paper introduces the process of 3-D reconstruction of surgery instruments based on the traditional method. It doesn't reconstruct the marks separately, but use the structure information as constraints of the projection functions to reconstruct the marks of one surgery instruments. The method can not only reduce the jitter greatly but also improve the accuracy effectively.3. Propose the prediction weighted filtering algorithm. It supposes the trajectory of marks and work point are quadratic curve. The coordinates in previous frame are used to fitting quadratic curve, with which predicted coordinates of current frame can be figured out. The final result is the weighted average of predicted and measured values. This algorithm, compared with common filtering algorithm, has a nice filtering effect while not affecting the real-time tracking. Based on the key technologies above, this paper develops the clinical prototype system of surgery navigation. The hardware includes binocular camera, instruments, focus objects and computer. The software is constituted by digital image acquisition, reconstruction of surgery instruments, reconstruction of medical image, as well as surgery navigation. The prototype system conducted two cases of clinical operation in the Shanghai Ninth People's Hospital, which indicated that its performance had reached the international advanced level.