Reduction of blocking artifacts by adaptive postfiltering of transform coefficients
Tao Chen, Hong Ren Wu, Bin Qiu · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2000
This paper proposes a novel postprocessing technique for reducing blocking artifacts in low bit rate transform coding. As the quantization is applied to transform coefficients in the encoding, the proposed approach works in the transform domain. Masking effect in the human visual system is considered, and an adaptive weighting mechanism is then integrated into the postfiltering. In low activity areas, since the blocking artifacts appear to be perceptually more detectable, a large window is used to efficiently smooth out the artifacts. In order to preserve image details, a small neighborhood and a large central weight are employed for processing those high activity blocks, where the blocking artifacts are less noticeable due to the masking of local background. The quantization constraint is finally applied to the postfiltered coefficients. Experimental results show that the proposed technique provides superior performance to other postprocessing methods in both objective and subjective image quality.