An improved algorithm for border following of binary images

Border following techniques have been extensively studied, and have a wide variety of applications. These existing methods can satisfactorily find and trace outermost borders. However, it is shown that interior holes are often incorrectly traced or follow sub-optimal paths. The authors examine the reasons for this, and put forward a new method that overcomes these deficiencies. Only the case of binary 2D raster images is examined. However, since such images form the basis for the more complex image types, the techniques described can be readily extended to the more complex image forms.<
>