In some examples, a system receives an input three-dimensional (3D) representation of a foot, and analyzes the input 3D representation of the foot to identify first points corresponding to metatarsal bones of the foot. Based on locations of the first points corresponding to the metatarsal bones, the system identifies a second point corresponding to a heel of the foot, and generates, based on the first points and the second point, a representation of features of the foot.