The skeleton of a digital image is a compact representation of its topology, geometry, and
scale. It has utility in many computer vision applications, such as image description,
segmentation, and registration. However, skeletonization has only seen limited use in
contemporary deep learning solutions. Most existing skeletonization algorithms are not
differentiable, making it impossible to integrate them with gradient-based optimization.
Compatible algorithms based on morphological operations and neural networks have been …

DRIVE The widely used DRIVE dataset consists of 40 twodimensional retinal color fundus
photographs and matching annotations of the visible blood vessels [9]. We normalize the
images to an intensity range between 0 and 1 and crop them to a size of 512× 512 pixel.
Then, we divide the dataset into training, validation and testing splits with a ratio of 60% to
20% to 20%. VesSAP The VesSAP dataset contains 24 three-dimensional light-sheet
microscopy images of murine brains after tissue clearing, staining, and labeling of the …