Local Natural Scale Based Contour Corner Detection Using Wavelet Transform

A new corner detection method for contour images is proposed based on dyadic wavelet transform (WT). The wavelet transform modulus maxima (WTMM) at different scales are taken as corner candidates. For each candidate, the scale at which the maximum value of the WTMM exists is defined as its local natural scale, and the corresponding modulus is taken as the significance measurement. This approach achieves more accurate estimations of the natural scale than the existing global natural scale methods. The simulation results show that the proposed method is effective for both long contours and short contours. The objective evaluation reveals the better performance of the proposed algorithm compared to the existing methods. The technique is inherently fast due to the fast implementations of the dyadic WT computations