Curvelet transform based adaptive image deblocking method

Blocking effect caused by coefficient quantification during image compression is an annoying problem. The main effect of quantification is to eliminate high frequency component in the image. Therefore, it will lead to noticeable discontinuous leaps which causes block effect and degrade the image. A deblocking method based on Curvelet transform is proposed in this paper. Based on the fact that the Curvelet coefficients in different scale layers correspond to different block effect of the degraded image, our method adaptively process the coefficients in each layer to recover the degraded images. As proved by the experiments, our method can retain more details and get better recovery results under both subjective and objective criterions than traditional spatial domain and wavelet deblocking methods.