医学影像后处理技术的研究及其在X线影像优化中的应用/

2019-04-20 22:45:00

image processing medical 影像 医学影像



现代医学离不开医学影像(图像)信息的支持。从伦琴发现X射线以来的一百多年,随着计算机、自动化、电子技术及其它技术的发展,人体成像的医学影像设备也不断创新,从二维的X线片到三维的CT重建图像,从反映人体解剖和结构形态的CT、MRI到揭示人体组织或器官功能、代谢的SPECT和PET。所有这些成像模式的产生,使得医学影像以及医学影像信息在临床诊断、治疗和医学研究中,正发挥着越来越重要的作用。如何更加充分地利用影像数据中所包含的可用信息,开发出对其它临床学科有用、更有诊疗价值的信息,以提高诊断效率和质量,是医学影像后处理技术研究的出发点和落脚点。
医学影像后处理技术的研究开始于70年代后期,随着影像设备的不断发展,这一领域的研究内容也在不断更新、发展、变化。近年来,由于计算机技术以及CT、 PET、MRI等技术的发展,特别是随着多媒体技术、网络技术的成熟及Internet的发展,对影像后处理提出的要求越来越高,其研究内容不断拓宽,并向纵深发展,主要包括影像增强、影像分割、影像配准与融合、影像压缩和可视化等方面。
医学影像后处理的首要任务就是影像增强,即对获取的医学影像进行增强信噪比的工作,滤除影像的噪声和干扰,突出感兴趣对象区域或边缘,为进一步分析和计算奠定基础。目前常用的增强方法包括对比度增强、平滑、锐化、伪彩色处理等。影像分割技术是对医学影像进行对象提取、定量分析、3D重建、配准融合等处理的一个必不可少的步骤。常用的影像分割技术包括阈值分割法、边缘检测法和区域分割法等等。影像配准与融合技术是近些年发展起来的后处理技术,经过配准、融合处理后的医学影像携带不同的生理、病理或解剖学方面的信息,对临床诊断和治疗具有极大的辅助作用。影像可视化技术将CT、MRI等数字化成像技术获得的人体信息在计算机上直观地表现为三维效果,能够弥补影像成像设备在成像上的不足,在辅助医生诊断、手术仿真、引导治疗等方面发挥了重要的作用。影像压缩技术是远程医疗和PACS系统发展的关键技术之一,减小医学影像占用存储空间、降低网络传输负担、提高压缩影像的质量、寻找有效的医学影像压缩编码是影像压缩技术的任务。
当前,在多种多样的医学影像中,各种X线影像数目占临床总图数的70%-80%,可以说X线影像是当前临床应用最广泛的医学影像。但是,由于人体结构和组织比较复杂,以及X线散射、电器噪声和光量子噪声等各种因素的影响,使得X线医学影像表现为动态范围宽、重叠度大、噪声高、细节丰富、数据量大和对比度差等特点。如何针对X线影像的特点,开发出有实际应用意义的影像后处理软件,以提取有用的病理信息,是提高当前医学水平的一个关键。
本文从实际应用出发,在深入研究医学影像增强、分割技术的基本原理及相关算法的基础上,针对目前使用量最大的X线影像,初步开发出了医学影像后处理系统应用软件。本系统结合了医学影像信息学与数字图像处理技术的相关知识,在开发过程中,主要进行了以下几个方面的研究:
1、医学影像后处理技术的基本概念及应用方向;
2、X线影像成像设备及原理,并在此基础上分析了X线影像的特点;
3、对医学影像增强技术、分割技术和配准与融合技术的基本原理及相关算法进行系统分析、研究。这是本论文研究的重点;
4、从系统框架、系统功能、开发平台、数据采集方式等方面对医学影像后处理系统进行详细分析。
在完成本论文的过程中,我所做的主要工作包括:
1、开发了医学影像后处理软件应用系统(主要针对X线影像),完成了对医学影像的对比度增强、平滑、锐化、伪彩色处理、固定阈值分割、自适应阈值分割和边缘检测等多种图像后处理和分析功能;
2、针对具有不同特点的X线影像,在验证各种算法的有效性的同时,选取其最优化处理方法;
3、为进一步开发影像配准和融合后处理功能模块,在技术理论上做了一些相关的预研准备。
本系统给医疗工作者提供了独立处理X线影像的手段,利用本系统提供的实用医学影像后处理功能,可以对X线影像进行优化处理,能够明显地改善X线影像的清晰度和图像判读的准确度。通过大量的实验证明,本系统具有明显的实用性,不仅提高了临床中的影像诊断水平,而且为医生制定合理的治疗方案提供了有力的支持。



Modern Clinical Medicine needs the supporting by medical images. A hundred years since the discovery of X-ray by Wilhelm Conrad Rontgen,the medical image equipment is innovated continuously with the development of computer, automatization, electronic technique and other techniques. From 2-D X-ray image to 3-D CT reconstructed image, from CT and MRI which can reflect human anatomy and configuration to SPECT and PET which can reveal the organization or the function and metabolism of organs of human body. All the imaging modalities make the medical images and image information playing a more and more important role in the field of clinical diagnosis and medicine research. The object of the medical image post-processing technique is how to make full use of the image information to improve the diagnostic efficiency and quality.
The research on medical image post-processing began from the evening of 70 years. With the development of the image equipment, the research of this field has been changing and improving. As a result of the development of computer science, CT, PET, MRI, especially the multi-media technique and Internet progress, of late years, the requirement is upsurging and going deeply. Now the medical image post-processing includes image enhancement, image segmentation, image registration and fusion, image compression and image visualization, ect.
The first task of the medical image post-processing is image enhancement, that is to enhance the Signal-to-Noise ratio, filtrate the noise, stand out the range of interests or edge and to lay a foundation for analysis and calculation. Contrast enhancement, smoothing, sharpening, pseudo-color processing are the methods in common. Image segmentation is absolutely necessary step for object separating, quantitative analyzing, 3-D reconstructing and image fusion, and so on. The methods used widely are threshold value segmentation, edge detection and region-based segmentation. Image registration and fusion developed in recent years. The medical image having been fused possesses the information from different physiology, pathology and anatomy, which are very valuable for clinical diagnosis. Image visualization can express the information of human body getting by CT, MRI into 3-D form, and remedy the shortcoming of imaging equipment. The technology plays an important role in diagnosis, surgery emulating and inducting treatment. Image compression is one of the key technologies for remote diagnosis and PACS. Its main goal is to decrease the storage room, reduce network transmission burden, improve the compressed image quality and looking for the high efficient compression code.
At present, in various medical images, the different kinds of X-ray images are widely used in clinical diagnosis. But influenced by the complexity of body tissues and some factors such as X-ray scatter and noise, the X-ray images display as the specialties of wide gray scale, high-degree overlap, high noises, large numbers of datum and low contrast. How to develop the applicable medical image post-processing system according the features of X-ray to extract the useful information, is the key point to improve the clinical diagnosis level.
From the point of application, this paper takes the living example of X-ray images used widely in most hospital, makes deep research on the base of the analysis and application of theory and technology of image enhancement, image segmentation and image registration and fusion, and develops the medical image post-processing system tentatively. The research combines medical image informatics with digital image processing, and the main studies as follow:
1. The basic conception and the application study of medical image post-processing;
2. Analyzing the factors of X-ray medical image based on the research of X-ray imaging equipment and theory;
3. In-depth research on the basic theory and the correlative algorithms of medical image enhancement and segmentation;
4. Analyzing the system frame, system functions, system running conditions, system development platform and data collection methods of medical image post-processing in detail.
The main work I did in the course of finishing this thesis is as follows:
1. Develops the medical image post-processing system, finishes the work of supporting contrast enhancement, smoothing, sharpening, pseudo-color processing, threshold value segmentation and edge detection, ect.
2. Validating the availability of algorithms and selecting the optimization processing method for different X-ray medical image.
3. Some research on medical image registration and fusion has been prepared for the development of the function Module.
The medical image post-processing system supports the functions of numerous medical image processing methods for medical operators. Post-processed by the methods, the quality of X-ray image is improved greatly. Plenty of experimentations testify that this system has the practicability. It increases the image diagnosis level and provides powerful sustain for reasonable medical treatment.