Fast Picture and Macroblock Level Adaptive Frame/Field Coding for H.264

Picture adaptive frame/field (PAFF) coding and macroblock adaptive frame/field (MBAFF) coding in the new video coding standard H.264 can improve compression efficiency for interlaced video coding. It is assumed that the optimal solution can be found by employing a multi-pass strategy, which encodes a macroblock (MB) or picture with all possible modes and chooses the one with smallest rate-distortion cost. However, this coding strategy needs high computational complexity. In this paper, a novel single-pass algorithm based on motion activity detection is proposed to predict the mode of PAFF and MBAFF coding in an analysis stage. The simulation results show that the proposed scheme can reduce computational complexity by approximate 60%~75% compared to the 3-pass encoder of reference software, while maintaining grate coding efficiency