眼科光学相干层析成像图像处理关键技术研究
文献类型：学位论文
| 作者 | 贺琪欲 |
| 学位类别 | 硕士 |
| 答辩日期 | 2016 |
| 授予单位 | 中国科学院上海光学精密机械研究所 |
| 导师 | 李中梁 |
| 关键词 | 光学相干层析成像 图像处理 视网膜 分层算法 伪彩色处理 |
| 其他题名 | Research on Key Technology of Image Processing for Ophthalmological Optical Coherence Tomography |
| 中文摘要 | 光学相干层析成像(Optical Coherence Tomography，OCT)为多种眼病特别是视网膜疾病的诊断提供了最直接的依据，对糖尿病视网膜病变、老年性黄斑变性等视网膜疾病的诊断具有重要甚至不可替代的作用，在眼科获得了广泛的临床应用。眼科OCT的原始图像仅是表征视网膜结构的灰度图像，而临床诊断需要观察更清晰、包含定量信息的图像。对眼科OCT原始图像进行去噪、图像增强、伪彩色等预处理，可使图像更清晰；对视网膜OCT图像自动分层并测量各层厚度，可提供疾病诊断所需的定量信息。眼科OCT图像处理技术可以有效地提高临床诊断的效率与准确性。本论文针对眼科OCT图像处理关键技术开展研究，主要工作包括以下两个方面： 1.提出了一种基于灰度特征的视网膜OCT图像自动分层方法。该方法利用三维块匹配与均值滤波相结合的散斑噪声抑制算法对视网膜OCT图像进行去噪预处理，然后使用分两步的分层算法对视网膜图像进行分层。分两步的分层算法首先在每个A-scan上设置可变阈值进行逐层分割作为初步分层结果，然后对各层的初步分层结果进行连续性和完整性判断和修正。采用该方法进行视网膜OCT图像的分层，可以精确地分出视网膜的9层结构；该方法能够适应高噪声和低对比度的图像，对存在血管等复杂结构的图像同样能够实现较好的分层。 2.提出了一种具有抑制散斑噪声功能的OCT图像伪彩色处理方法。该方法首先使用多级阈值将灰度图像转换为索引图像，然后确定多级阈值中对应图像散斑噪声的级次，再创建特定的颜色表，并使用创建的颜色表来显示索引图像，从而实现对灰度OCT图像的伪彩色处理并抑制原图中的散斑噪声。采用上述方法可以对高噪声、低对比度的OCT图像进行快速、良好的伪彩色处理，伪彩色处理结果有效地削弱了散斑噪声的影响，增强了图像的对比度，改善了显示效果。 |
| 英文摘要 | Optical coherence tomography (OCT) provides the most direct basis for the diagnosis of many eye diseases, especially retinal diseases. It plays an important and even irreplaceable role in diagnosis of retinal diseases such as diabetic retinopathy and age-related macular degeneration, and it has been widely used in ophthalmology. The original images of ophthalmological OCT are merely gray scale images displaying the retinal structure. However, more clear images containing quantitative information is needed in clinical diagnosis. The images can be made more clear by applying preprocessing such as image denoising, image enhancement and pseudocolor to the original images of ophthalmological OCT. Automatic segmentation of retinal OCT images and thickness measurement of each retinal layer provides quantitative information for the diagnosis of eye diseases. The application of image processing technique in ophthalmological OCT can effectively improve the efficiency and accuracy of clinical diagnosis. Key technology of image processing for ophthalmological OCT is studied in this dissertation, and the main researches are as follows: 1. An automated retinal layer segmentation method based on intensity feature is proposed. A two-step search that employs block-matching and 3D filtering along with mean filtering is used as a preprocess in the method. The two-step search begins with segmenting individual retinal layers by setting a variable threshold on each A-scan as initial results, which are then checked and corrected for continuity and integrity. The experimental results show that the proposed method provides accurate segmentation of nine retinal layers and it can be applied to OCT images affected by speckle noise, low image contrast, and even with irregularly shaped structural features such as blood vessels. 2. A pseudocolor method that can suppress speckle noise for OCT images is proposed. In this method, the gray scale image is firstly transformed into an index image by using multilevel threshold. The gradations of multilevel threshold that correspond to speckle noise in the image are then determined and specific color map is created. The created color map is used to display the index image so as to realize the pseudocolor processing method and suppress speckle noise in the original image. The proposed pseudocolor method is fast and effective, and can be applied to OCT images with high noise and low contrast. The results show that the method effectively suppresses the speckle noise, enhances the image contrast and improves the display effect. |
| 语种 | 中文 |
| 源URL | [http://ir.siom.ac.cn/handle/181231/16971]  |
| 专题 | 上海光学精密机械研究所_学位论文 |
| 推荐引用方式 GB/T 7714 | 贺琪欲. 眼科光学相干层析成像图像处理关键技术研究[D]. 中国科学院上海光学精密机械研究所. 2016. |
入库方式： OAI收割
来源：上海光学精密机械研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。