Structural Feature Extraction from Regular-Shaped Rigid Objects(1) - Based on Gradient Direction Constraint Randomized Hough Transform

Structural Feature Extraction from Regular-Shaped Rigid Objects(2) - Enhanced by Prior Knowledge

“…(c) (d) Fig.4 Results of corner extraction enhanced by the prior knowledge-based rules same subject (1). The numbers in Fig.4 denote the label of every connectivity component of the corner features, which is different from Fig.3.…”

“…Our focus is on the edges) as the white arrow points can be recovered by the intersection of the prolonged straight edges/line segments of No.3 and No.4 corners. In our work, as the same subject (1) shows, the straight edges are only extracted in the local windows. To reach the other corner, a local straight line segments should be prolonged along the relative whole straight edge until the end of the edge.…”

“…As for the second rule, 1 2 | | m m m t (1) is used where denote the slopes of every two straight edges in the local coordinate system. (2)…”

“…Structural feature extraction has been studied in the previous report of the same subject [1] (In the sequel, that paper will be stated as "the same subject (1)") by the same authors. However, as the same subject (1) says, its results still have some problems from the viewpoint of recognition.…”

“…The limitation of the same subject (1) is that it relies only on the low-level image data. As stated in the same subject(1), structural recognition is a humanoid manner, hence knowledge information can be and should be exploited.…”