Figure 5
From: Integrated deep learning framework for accelerated optical coherence tomography angiography
Qualitative performance of our DenseReconstGAN and HARNet for reconstructing at a downsampling ratio of two. Enlarged profiles of the boxed regions of interest (ROI) are illustrated. Vessel profiles along the dashed lines within each ROI are also illustrated. (LR, LQ) low-resolution and low-quality, (HR, HQ) high-resolution and high quality.
