Locating perceptually salient points on planar curves

Corner Detection and Curve Partitioning Using Arc-Chord Distance

“…The work of Fischler and Wolf [9] extends the technique of [3]. An important contribution of their work over [3] is a major revision of the approach to filtering the critical points, based on comparisons at a given scale as well as across different scales (i.e., different values of the input parameter k).…”

“…A critical problem in machine vision is how to breakup (partition) the perceived world into coherent or meaningful parts prior to knowing the identity of these parts [9]. In particular, the problem of partitioning digital planar curves has been a subject of intense investigation since the earliest days of machine vision.…”

Estimating Angles And Curvature Features In Grey Scale Images

“…In classic mathematics, curvature is defined as the rate of the angle change along the arc length. Although some researchers argue that saliency in digitised image can be best captured by other means [5], curvature remains to be a generic, concept of saliency on curves. The major challenge of curvature based approach is then how to extract the curvature feature points in the presence of noise accurately.…”

Localization on curved objects using tactile information

“…The algorithm converges to the first feasible segment satisfying (4) and (5) in the marching direction.…”

“…Fischler [4] described an algorithm to locate points with extreme curvatures on planar curves and reconstruct the original curves based on these points. Mokhtarian and Mackworth [9] used inflection points (where curvatures are zero) for planar curve descriptions and matching.…”