Fig. 4
From: Coronary artery segmentation in angiographic videos utilizing spatial-temporal information
Examples of blood vessel segmentation results. Row 1 ∼5 correspond to five target images captured in five video clips, respectively. Column A: target images containing coronary arteries; col. B: ground truths, manually labeled blood vessels; col. C: segmentation results of U–Net; col.D: results of AG–Net; col. E: results of 2D CE–Net; col. F: results of our 3D–2D model, where N is 1; col. G: result of 3D–2D model, where N is 2