【作者】 杨丹军；

【导师】 王保保；

【作者基本信息】 西安电子科技大学 ， 计算机应用技术， 2010， 硕士

【摘要】 计算机辅助诊断是最近十几年发展起来的生物医学工程的一个重要分支。随着数字图像技术的发展,计算机辅助消化道内窥镜图像诊断成为可能。其充分利用合适的数字图像处理技术来分析处理医学图像信息,对组织结构信息进行定性和定量研究,从而辅助医生进行诊断治疗。本课题的研究目的是提出了一种新的计算机辅助消化道内窥镜图像诊断方法,用于消化道内窥镜图像的自动分析。本文使用了一种基于多级同质的彩色消化道内窥镜图像分割方法。该方法首先计算出关于强度和同质的二维直方图,然后在二维直方图上利用峰值寻找算法将图像分割成若干个区域,接着利用色调域直方图分析法将前面所得的每个区域进行再分割,最后合并具有相同CIE(Lab)颜色度量的区域。本文对彩色消化道内窥镜图像提取了基于直方图的颜色信息和基于纹理谱的纹理信息。综合颜色和纹理的图像特征将具有更强的鲁棒性。Bayesian分类方法是一种常用的统计模式识别方法。本文将其首次应用于消化道内窥镜图像的诊断,并获得了较好的实验结果。针对息肉的检测,本文提出了一种新的基于椭圆匹配的息肉检测算法。该算法首先对消化道内窥镜图像进行彩色边缘检测并最终处理成二值边界。接着用一种基于最小二乘法的随机化椭圆检测方法在二值边界图像中进行椭圆检测。实验取得了很好的效果,并弥补了Bayesian分类方法不能识别息肉的缺点。实验结果初步显示了论文所提出的方法在消化道内窥镜图像诊断上的可行性,有望为消化道疾病的自动诊断提供一种新的分析方法。更多还原

【Abstract】 Computer-aided diagnosis is developed in recent decades, which is an important branch of bio-medical engineering. With the development of digital image technology, computer-aided diagnosis of digestive endoscopic image becomes possible. It makes full use of appropriate digital image processing techniques to analyse medical image and gets qualitative and quantitative analysis of the tissue, which can facilitate the diagnosis and treatment of doctors. The purpose of this paper is to propose a new method of computer-aided diagnosis of digestive endoscopic image for the automatic analysis of digestive endoscopic image.This paper uses a hierarchical approach to color digestive endoscopic image segmentation using homogeneity. In the first stage, the regions are segmented using a peak-finding algorithm on a 2-D histogram of homogeneity and intensity values. In the second stage, histogram analysis of the color feature hue is performed to subdivide the segmented regions obtained from the first stage. The subdivisions of different segmented regions having similar CIE(L*a*b) color measure are merged.In this paper, both color-based and texture-based quantitative features of color digestive endoscopic image are extracted. Specifically, we extract texture-based features from texture spectra and color-based features from color histogram. Integrated color and texture features of the image will have a stronger robustness.Bayesian classifier is one of the most commonly used method of statistical pattern recognition. This paper uses it to the diagnosis of digestive endoscopy image for the first time and obtains good results.For polyp detection, this paper proposes a novel scheme based on ellipse fitting for polyp detection in digestive endoscopic image. Firstly, a color edge detection algorithm is used to get the binary image. Then, we adopt a randomized ellipse detection algorithm based on the least square approach to detect polyps in the binary image. Experiment has achieved very good results and the algorithm can overcome the shortcoming of Bayesian classifier’s missing detection of polyps.Experiment results suggest the feasibility of the proposed method for the diagnosis of digestive endoscopic image, which provides a new analysis method for the automatic diagnosis of digestive endoscopic image.更多还原