Reduction of blocking artifacts by cepstral filtering

We propose a convolution model for the blocking artifacts in compressed images and videos, and a cepstral filtering algorithm for reducing the artifacts. Since each of the horizontal or vertical line of the compressed image has abrupt changes at almost every 8th samples, it can be modeled as a convolution of the original sequence with an excitation signal of periodicity 8. Comparison of the cepstrums of the compressed and original images shows that the multiples of 8th samples of the compressed signal’s cepstrum usually have larger magnitudes than those of the original’s, which supports the proposed model. The proposed filtering algorithm is to reduce the magnitudes of every 8th samples of the cepstrum, and disperse the reduced values to the neighboring samples. The algorithm is conducted in the cepstrum domain for reducing only the excitation signal which is the cause of blocking artifacts. Hence, the high frequency contents are not reduced by the postprocessing, whereas the conventional lowpass filtering blurs the image.