多尺度二维直方图均衡化算法在医学图像增强中的应用研究

发布时间：2018-05-17 14:55

本文选题：高比特医学图像 + 图像增强　；参考：《东北师范大学》2017年硕士论文

【摘要】：医学成像技术已经逐渐成为医疗诊断过程中对人体结构进行无损检测的重要工具。它主要包括X射线成像、计算机断层扫描成像(CT)、磁共振成像(MRI)、超声成像和核医学成像等。这几种成像方式分别有其独特的应用领域,针对不同的生理结构或病理情况,选用相应的医学成像技术。作为健康检查和重大疾病检测成像技术的医学CT图像和MRI图像,可以展示被测人体的内部结构和状态,因此,图像质量的好坏尤为重要。DICOM是医学图像的国际标准,被广泛应用于医学成像领域,是一种高比特医学图像格式,具有信息量大,内容丰富等特点。由于成像设备获得的原始图像受到设备本身硬件性能的制约和获取条件等多种因素的影响,直接从医学仪器获得的图像会出现图像降质现象,比如对比度较低、图像模糊等。针对这一情况,本文以高比特医学图像作为处理对象,提出一种新的基于Laplacian金字塔和二维直方图均衡化(2DHE)相结合的多尺度二维直方图均衡化算法,简称为Lap-2DHE。在该算法中,引入Laplacian金字塔将图像层分离,基于图像灰度统计分布自适应选取参考点调整塔层系数,从而对图像进行多尺度分析,以增强图像细节。再结合二维直方图均衡化(2DHE),并采用基于模糊信噪比BSNR的方法来控制邻域大小的选择,使图像增强达到最佳效果。本文算法有效解决了传统2DHE方法在高比特医学图像增强过程中细节容易丢失的问题。实验结果表明,对比已有算法,Lap-2DHE算法能有效的提升对比度,增强图像细节,改善图像视觉效果,对医学图像增强具有普遍适用性,对于提升无损检测技术的实用性具有重要意义。通过图像增强算法提升医学图像质量,不仅提高了医疗设备实用性和可靠性,还打开了基于低剂量电磁波的成像技术应用的局面,在医学检测中具有很好的应用前景,为医疗成像设备的发展和应用提供了可行性支持。
[Abstract]:Medical imaging technology has gradually become an important tool for nondestructive testing of human structure in the process of medical diagnosis. It mainly includes X-ray imaging, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound imaging and nuclear medicine imaging. These imaging methods have their own unique application fields, according to different physiological structure or pathological situation, the corresponding medical imaging technology is selected. Medical CT images and MRI images, as health examination and major disease imaging techniques, can display the internal structure and state of the human body under test. Therefore, the quality of images is particularly important. DICOM is the international standard of medical images. It is widely used in the field of medical imaging, is a high bit medical image format, with a large amount of information, rich content and so on. Because the original image obtained by the imaging equipment is affected by many factors such as the hardware performance of the equipment and the condition of obtaining, the image obtained directly from the medical instrument will appear the phenomenon of image degradation, such as low contrast, image blur and so on. In this paper, a new multi-scale two-dimensional histogram equalization algorithm based on Laplacian pyramid and two-dimensional histogram equalization is proposed, which is called Lap-2DHE. In this algorithm, the Laplacian pyramid is introduced to separate the image layer, and the reference point is selected to adjust the tower layer coefficient adaptively based on the image grayscale statistical distribution, so that the multi-scale analysis of the image is carried out to enhance the details of the image. Combined with two-dimensional histogram equalization and fuzzy signal-to-noise ratio (BSNR) method to control the selection of neighborhood size, the image enhancement can achieve the best results. This algorithm effectively solves the problem that the details of traditional 2DHE method are easily lost in the process of high bit medical image enhancement. The experimental results show that compared with the existing Lap-2DHE algorithm, it can effectively enhance the contrast, enhance the details of the image, improve the visual effect of the image, and have universal applicability for medical image enhancement. It is of great significance to improve the practicability of nondestructive testing technology. Improving the quality of medical image by image enhancement algorithm not only improves the practicability and reliability of medical equipment, but also opens up the application of imaging technology based on low dose electromagnetic wave, which has a good application prospect in medical detection. It provides feasibility support for the development and application of medical imaging equipment.
【学位授予单位】：东北师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318;TP391.41

【参考文献】