CASIA OpenIR  > 毕业生  > 硕士学位论文
基于X-射线影像的手指骨图像增强算法研究
Alternative TitleImage Contrast Enhancement Based on X-Ray Phalange Images
陈亿霖
Subtype工学硕士
Thesis Advisor蒋田仔
2004-06-11
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword手骨图像 图像增强 同态滤波器 Radon变换 Curvelet变换 关联系数 自动评价系统 Phalange Image Skeletal Maturity Assessment Image Contrast Enhancement Homomorphic Filter Radon Transform Curvelet Transform Ass
Abstract骨发育成熟度(骨龄)自动评判技术的出现得益于现代图像处理技术和模 式识别技术的发展。骨龄指标在预防医学、临床医学和体育科学等领域得到广 泛的应用,并且日益受到关注。由于准确性和客观性的要求,世界各国均大力 发展本国的骨龄自动评判系统,近年来更是取得了长足的进步,得到了大量喜 人的技术突破。骨龄识别大致可以分为四个阶段:手骨图像的预处理、图像的 增强、图像特征的提取和图像特征匹配。在本文中,我们有针对性的对骨龄评 价的前两个阶段进行了算法研究,提出了自己的研究观点,并且致力开发出骨 龄评判的自动评价系统。 本文针对骨龄系统的图像增强环节进行了初步的探索性研究,涉及了许多 图像处理的基本问题,包括图像的滤波、图像的增强、图像的傅里叶变换、图 像的Radon变换、图像的小波变换和图像的Curvelet变换等。在本文中,主要 的工作和贡献有: I.提出并实现了一个基于频域滤波的同态滤波器。在实现过程中,我们综 合了频域低通滤波器和频域高通滤波器的特点,对同态滤波器进行了改 进,并且从“振铃效应”、“图像模糊程度”和‘‘噪声平滑效果”这三个 方面与其他的经典滤波器进行了比较,取得了较好的效果。 2.针对图像的特点,我们通过分析选择使用了Curvelet的方法进行图像增 强。首先,我们分析了原来Radon离散算法的一些不足,它会由于采 样的不足导致图像重建以后在某些边缘出现“环绕现象”。对此,我们 提出了改进方法,即首先对其进行图像沿展,再进行计算,最后在重建 之后的图像中进行截取,得到所需图像。 3.改进了原来的修改Curvelet系数的算法。对于原来的修改Curvelet系数 的算法,由于其固有原理的原因,使得有些信号被误判为噪声而被抹去。 我们在这里引入“关联系数”,通过关联系数和原算法相结合的方法共 同修改和选取Curvelet系数,并最后根据这些系数进行重建,改进的方 法对于细节的描述和表达比原算法有了一定的提高。4.根据以后分割的需要和进行特征匹配的需要,对图像进行后处理,例如 姿态调整和手腕部截取等。 总得说来,本文在针对手骨图像的图像增强环节进行了有益的探索。
Other AbstractAccurate and objective skeletal maturity (i.e. bone age) assessment becomes more and more important to the realm of prophylaxis, therapeutics, and athletic sports, and so on. Traditional technologies cannot satisfy the increasing requirements. Recently, the technology of automated assessment has made a great deal of progress and got much promising technical breakthrough. Bone age assessment can be divided into four periods: preprocessing of X-ray phalange image, image contrast enhancement, feature extraction and feature matching. This thesis focuses on the image contrast enhancement. The main contributions are as follows: We propose and implement a homomorphic filter based on the frequency filtering. In the implement procedure, I incorporate the characteristics of the frequency domain lowpass filter with the frequency domain highpass filter, and then design a modified homomorphic filter. For the image oriented some special applications, we develop a fast radon discrete transform algorithm. After a variety of experiments, I choose curvelet algorithm for the image contrast enhancement. As for the "low sampling rate" of the radon discrete transform, the reconstructed images show the "wreathed effect", I firstly expand the image, then compute the curvelet coefficient, lastly intercept the reconstructed image of antieipant size. By the means mentioned above, the "wreathed effect" disappears. We present and propose an algorithm used for improving the curvelet coefficients. I introduce the concept of associate coefficient which is widely used in wavelet, and combine the concept with the traditional methods to modify the curvelet coefficients, and then propose an integrated method.
shelfnumXWLW786
Other Identifier786
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/6751
Collection毕业生_硕士学位论文
Recommended Citation
GB/T 7714
陈亿霖. 基于X-射线影像的手指骨图像增强算法研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2004.
Files in This Item:
File Name/Size DocType Version Access License
硕士生学位论文-786.pdf(9298KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[陈亿霖]'s Articles
Baidu academic
Similar articles in Baidu academic
[陈亿霖]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[陈亿霖]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.