A low bit-rate video coding approach using modified adaptive warping and long-term spatial memory

Abstract

In this paper, an H.264/AVC video coding strategy is introduced that employs a spatial-temporal video sequence representation in which video frames are coded at a low spatial sampling rate and reference I frames are coded at high spatial resolution. High spatial frequency information is re-synthesized at the receiver side using an adaptive motion estimation and warping method. The approach as presented is shown to improve coding quality for sequences with low to moderate motion.

An approach to enhanced definition video coding using adaptive warping

“…The rest of the frames are coded as P frames. As reported by Chen et al, 6 the forward warping algorithm will fail on sequences with moderate to high motion because the motion can no longer be accurately represented with translational pixels. Depending on the nature of the motion, as shown in Figure 8 where it depicts the 28th frame in the Stefan sequence, the original forward warping algorithm produces false motion fields, leading to a distorted area around the tennis player's left arm.…”

“…As was reported by Chen et al, 6 when the motion is high in the sequence and the target frame is sufficiently dissimilar to the reference frame, warping distortion can occur. The degradation does not show up as a lack of sharpness but rather as a twisting of contours and edge boundaries.…”

“…After being transmitted and decompressed, the video frames are expanded back to the original size using an adaptive warping method and previously decoded high spatial definition I-frames. 6 This approach was shown to work effectively when motion was not too severe. In high motion situations, however, the reconstruction can and often does incur warping errors, which are visually unacceptable.…”