基于分类器技术的白内障组织图像去噪方法

Cataract organization image denoising methods based on classifier technology

为实现根据视频图像等信息自动识别白内障组织及硬度,并对其进行智能分类的目的,提出了基于分类器技术的白内障组织图像去噪方法。对白内障图像去噪的关键在于区分白内障与正常组织和区分白内障硬度级别。这两个方面均涉及了图像特征提取和分类器方法。对图像特征提取,采取了颜色特征和纹理特征进行提取,并在这两种传统的特征上提出了一种新的图像特征提取方法。在分类器算法的选择上,采取了最近邻法和支持向量机的方法。以某医院提供的手术视频截图为例,对比了采用不同方法对白内障组织的识别率。

To automatically identify the cataract tissue,including the hardness and intelligent classification,cataract organization image denoising methods based on the classifier technology is presented.The key issues to implement cataract image denoising are distinguishing cataract with normal organization and distinguishing cataract hardness level.Both of the two aspects are related to the image feature extraction and classification methods.The color feature and texture feature are adopted on the image feature extraction,and a new image feature extraction method is put forward based on these two traditional features.The nearest neighbor method and the support vector machine are used as classifier algorithm.Finally,a hospital surgical video capture is taken as example to demonstrate the recognition rate of cataract organization by using different methods.