Scholarlink

<p>Optical coherence tomography angiography (OCTA) is a non-invasive imaging technique, and automatic segmentation of retinal vessels is crucial for understanding ocular diseases and making informed clinical decisions. However, the automatic segmentation of retinal vessels in OCTA images is particularly challenging due to several inherent issues. Retinal vessels often exhibit low contrast against the surrounding tissue, making it difficult to distinguish them clearly. Additionally, the complex and irregular branching structures of retinal vessels, along with the presence of noise and artefacts in OCTA images, further complicate the segmentation task. To address these challenges, we propose a novel method, the spatial and multi-frequency domain-based segmentation network (SMFDNet), specifically designed for vessel segmentation in OCTA fundus images. This network effectively combines spatial and multi-frequency domain features to enhance the segmentation accuracy of retinal vessels. To demonstrate the superiority of our proposed network, we conduct experiments on the Retinal Vessels Images in OCTA (REVIO), Retinal OCT-Angiography Vessel Segmentation (ROSE) and Optical Coherence Tomography Angiography-500 (OCTA-500) datasets. The extensive experimental results show that our approach consistently outperforms state-of-the-art methods, particularly in handling low contrast and complex vessel structures. The codes are available at <a href="https://kyanbis.github.io/SMFDNet/">https://kyanbis.github.io/SMFDNet/</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

SMFDNet: spatial and multi-frequency domain network for OCT angiography retinal vessel segmentation

  • 期刊论文

  • 发表日期：

    2025年2月24日

查看全文
  • Sien Li,
  • Fei Ma,
  • Fen Yan,
  • Jing Meng,
  • Yanfei Guo,
  • Hongjuan Liu,
  • Ronghua Cheng

摘要

Optical coherence tomography angiography (OCTA) is a non-invasive imaging technique, and automatic segmentation of retinal vessels is crucial for understanding ocular diseases and making informed clinical decisions. However, the automatic segmentation of retinal vessels in OCTA images is particularly challenging due to several inherent issues. Retinal vessels often exhibit low contrast against the surrounding tissue, making it difficult to distinguish them clearly. Additionally, the complex and irregular branching structures of retinal vessels, along with the presence of noise and artefacts in OCTA images, further complicate the segmentation task. To address these challenges, we propose a novel method, the spatial and multi-frequency domain-based segmentation network (SMFDNet), specifically designed for vessel segmentation in OCTA fundus images. This network effectively combines spatial and multi-frequency domain features to enhance the segmentation accuracy of retinal vessels. To demonstrate the superiority of our proposed network, we conduct experiments on the Retinal Vessels Images in OCTA (REVIO), Retinal OCT-Angiography Vessel Segmentation (ROSE) and Optical Coherence Tomography Angiography-500 (OCTA-500) datasets. The extensive experimental results show that our approach consistently outperforms state-of-the-art methods, particularly in handling low contrast and complex vessel structures. The codes are available at https://kyanbis.github.io/SMFDNet/.