A new approach for reducing blockiness in DCT image coders

This paper presents a new approach for reducing the blockiness that occurs when using DCT image coders at high compression ratios. The method is simply the replacement of the inverse DCT-2 in the decoder by a larger inverse DCT-1 followed by overlapping and averaging of the enlarged blocks to reconstruct the image. The modified decoder can decode any bitstream generated by a standard encoder. Blockiness is reduced but there is no noticeable distortion or loss of sharpness in the image. There is also no significant increase in complexity when using this method.