Figure 10From: Modeling sets of unordered points using highly eccentric ellipsesThe original image, manual segmentation, data used in our retinal parsing algorithm, and confidence regions. (a) The original retinal image. (b) The manual segmentation of the vessels in (a). (c) Result of thinning the image in (b). (d) The confidence regions depicted as circles with a radius equal to 1% of the diagonal of the bounding box of the original set. The figure is better seen in color.Back to article page