| 基于图像生成的多模态视网膜图像配准方法 |
| |
| 引用本文： | 於佳乐,黄 坤,张 潇,陈 强.基于图像生成的多模态视网膜图像配准方法[J].南京师范大学学报,2023,0(1):010-17. |
| |
| 作者姓名： | 於佳乐 黄 坤 张 潇 陈 强 |
| |
| 作者单位： | (南京理工大学计算机科学与工程学院,江苏 南京 210094) |
| |
| 摘 要： | 针对多模态视网膜的全局粗配准,提出了一种基于图像生成的多模态视网膜配准方法. 不同于当前主流的提取多模态视网膜图像血管再进行配准的方法,使用GAN模型对不同模态的视网膜图像进行像素级映射,再通过特征点匹配的方式计算图像仿射矩阵,实现图像粗配准. 基于彩色眼底图像与荧光素血管造影图像的实验结果表明,该方法与当前主流的视网膜粗配准方法相比,具有快速且鲁棒的优势.
|
| 关 键 词： | 图像配准 多模态 视网膜图像 图像生成 |
Multi-Modal Retinal Image Registration Method Based on Image Generation |
| |
| Yu Jiale,Huang Kun,Zhang Xiao,Chen Qiang.Multi-Modal Retinal Image Registration Method Based on Image Generation[J].Journal of Nanjing Nor Univ: Eng and Technol,2023,0(1):010-17. |
| |
| Authors: | Yu Jiale Huang Kun Zhang Xiao Chen Qiang |
| |
| Affiliation: | (School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China) |
| |
| Abstract: | A multi-modal retinal registration method based on image generation is proposed for global coarse registration of multi-modal retinal images. Unlike the current mainstream methods that extract retinal vascular structures for registration, this method uses GAN model to perform pixel-level mapping of different modal retinal images. Then, the affine matrix is calculated through feature point matching to achieve image rough registration. Experimental results based on color fundus images and fluorescein angiography images demonstrate that this method has the advantages of faster speed and robust performance compared with current mainstream retinal registration methods. |
| |
| Keywords: | image registration multi-modal retinal images image generation |
|
| 点击此处可从《南京师范大学学报》浏览原始摘要信息 |
|
点击此处可从《南京师范大学学报》下载全文 |
|
