Previous Works

OCTA Capillary Segmentation

The images displayed here have been precisely segmented using an advanced AI model tailored for OCTA capillary analysis (masks on the right side are painted on top of the original image). For both training the model and showcasing the results, we utilized the comprehensive OCTA-500 dataset [1,2,3].

OCTA Image
  1. Mingchao Li, Yerui Chen, Zexuan Ji, Keren Xie, Songtao Yuan, Qiang Chen, and Shuo Li. "Image Projection Network: 3D to 2D Image Segmentation in OCTA Images," IEEE Trans. Med. Imaging, vol. 39, no. 11, pp. 3343-3354, 2020.
  2. Mingchao Li, Weiwei Zhang, and Qiang Chen. "Image Magnification Network for Vessel Segmentation in OCTA Images," in Chinese Conference on Pattern Recognition and Computer Vision. arXiv:2110.13428, 2022.
  3. Mingchao Li, Kun Huang, Zetian Zhang, Xiao Ma, and Qiang Chen. "Label Adversarial Learning for Skeleton-Level to Pixel-Level Adjustable Vessel Segmentation," arXiv:2205.03646, 2022.

Fluorescein Angiography Segmentation

The images below show how the AI-based image segmentation works on fluorescein angiography (FA) images. On the left hand side there is an original image, and on the right is the image with the segmentation mask painted on top. The model is developed on the FA-recovery19 dataset [1].

FA Image
  1. Li Ding, Mohammad H. Bawany, Ajay E. Kuriyan, Rajeev S. Ramchandran, Charles C. Wykoff, Gaurav Sharma, June 3, 2019, "RECOVERY-FA19: Ultra-Widefield Fluorescein Angiography Vessel Detection Dataset", IEEE Dataport, doi: https://dx.doi.org/10.21227/m9yw-xs04.