A method for removing blocking effects in MPEG-2 video by applying a block classification technique using stream information

This paper proposes a method for removing the blocking effects that become evident when decoding MPEG-2 video data. The method classifies each block and detects blocking effects by using the information obtained from the MPEG-2 video bitstream, including DCT coefficients and motion vectors. Blocking effect removal is achieved by applying a filter to the decoded images in response to the detected result. The effectiveness of the method has been verified by computer simulation. Based on subjective judgment results, the image quality encoded at 3 Mbps by using this method was confirmed to be closely comparable with the quality of images encoded at 4 Mbps. Because this method carries out noise elimination by only using the information obtained from the MPEG-2 video bitstream, it can be applied to any application, including storage, broadcasting, communications, etc.