Abstract:
The spatial domain, frequency domain and wavelet based techniques are
being used independently to detect edges in an image. Spatial filters are very
good at localization accuracy but do not have any control over the operator’s
scale. Similarly Fourier transform being global in nature can neither localize sharp
transients nor differentiate between true and false edges. The problem
aggravates further under noisy scenario. The classical edge detectors do not yield
adequate edge maps of the noisy images and malfunction over default threshold
values. The choice of optimum threshold for edge detection is not generic for
diverse set of images and noise models. A good threshold assigned to yield a good
edge map for a particular type of image and noise model may be inappropriate
for other type of image or the different noise model/intensity in the image. Thus
it requires user’s intervention to assign suitable threshold value to differentiate
between true and false edges. Thus the two major dilemmas for edge detection
are firstly the choice of appropriate threshold to segregate noise and true edges
and secondly to opt for an appropriate scale for edge detection. In this research work a novel edge detection paradigm is envisaged to work for various images. The image is decomposed by multilevel wavelet decomposition using Quadrature mirror filters and then thresholded. The decomposition level is determined by the image resolution. The property that image structural details remain present at each level whereas noise is partially eliminated within subbands, is being exploited. The lower resolution wavelet detail bands are interpolated to theoriginal image size which partially recaptures the missing edge pixels besidesfacilitating matrix multiplications. An innovative wavelet synthesis approach is conceived based on wavelet scale correlation of the concordant detail bands such that the reconstructed image fabricates an edge map of the image.
