Edge Detection in Digital Images Using Dispersive Phase Stretch Transform

Figure 2

Effect of warp () and the strength () of phase applied to a sample image (shown in (i)) on the edge detection performance in the proposed method. (a) to (f) Comparison of edge detection results with three different amounts of in the phase applied to the image. Phase derivative profiles are shown in (a) to (c) and the corresponding edge detection results are shown in (d) to (f). As seen, edge detection with medium warp has better noise performance than the case with very large warp or the case with linear phase derivative; compare regions indicated with white triangular, circle, and rectangular markers. (e), (g), and (h) Comparison of edge detection performance for the case of with three different amounts of applied to the image. As seen, larger phase results in less edge noises but also less resolution to detect the edges. Thus, there is an optimum value for and parameters to reduce the edge noises while preserving the vital edge information.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i) Original image