Phase Congruency Based Denoising of Historical Document Images

When Fourier component of image are maximally in phase then image features like edges, lines and match bands are present at that point. For detecting edges of image phase based method is more suitable than gradient based method because the phase is dimensionless quantity. Change in phase will not change the brightness or contrast of image. This provides a threshold value which can be applied over an image and binarized image is obtained. Here phase congruency features are calculated using wavelets. This paper provides the information about use of non-orthogonal, complex valued, log-Gabor wavelets, rather than the more usual orthogonal or bi-orthogonal wavelets. Thresholding of wavelet responses in the complex domain allows one to ensure that perceptually important phase information in the image is not corrupted. It is also shown how appropriate threshold values can be determined automatically from t1he statistics of the wavelet responses to the image