术中白内障组织碎片硬度识别的分类器技术研究

Classifier Technology for Tissue Fragment Hardness of Cataract Intraoperative Recognition

为了根据图像和其它相关信息快速识别超声乳化针头所触及到的是正常组织还是白内障组织,如果是白内障组织,则需要进一步识别其硬度,为能量释放模式的控制提供依据,提出一种使用最近邻法分类器、人工神经网络分类器以及多分类器集成等方法对白内障组织碎片图像进行分析,对其硬度进行判别归类。该方法不仅能实现白内障组织碎片硬度的自动识别、改善白内障手术效果,而且能极大降低白内障手术的难度,减少了人为误操作给患者带来的不必要损伤。

In order to identify if the fragment which ultrasonic emulsification needle touched is normal tissue or cataract according to the images and other relevant information, and if it is the cataract to identify it＇ s hardness to provide the basis for the control of energy release model, this paper put forward some integration methods using the nearest neighbor classifier, artificial neural network classifier and multiple classifier to analyze tissue images of cata- ract fragments and classify their hardnesses. This method can not only realize cataract fragments, improve the cataract surgery effects, but also greatly reduce the difficulty of cataract surgery, reduce unnecessary damages of patients brought by man-made operations.