Dot to dot, simple or sophisticated: a survey on shape reconstruction algorithms

Dot pattern points are the samples taken from all regions of a 2D object, either inside or the boundary. Given a set of dot pattern points in the plane, the shape reconstruction problem seeks to find the boundaries of the points. These boundaries are not mathematically well-defined. Hence, a superior algorithm is the one which produces the result closest to the human visual perception. There are different challenges in designing these algorithms, such as the independence from human supervision, and the ability to detect multiple components, holes and sharp corners. In this paper, we present a thorough review on the rich body of research in shape reconstruction, classify the ideas behind the algorithms, and highlight their pros and cons. Moreover, to overcome the barriers of implementing these algorithms, we provide an integrated application to visualize the outputs of the prominent algorithms for further comparison.