Video prediction block structure and the emerging high efficiency video coding standard

In the ISO/IEC 14496-10 |ITU-T H.264 advanced video coding (AVC) standard, the prediction block sizes can be 16×16, 8×8 and 4×4 for Intra prediction; 16×16, 16×8, 8×16, 8×8, 8×4, 4×8 and 4×4 for Inter prediction. In the first HEVC test model (HM1.0), each 2N×2N Intra CU may consist of either one 2N×2N prediction unit (PU) or four N×N prediction units; while each 2N×2N Inter CU may consist of one 2N×2N PU, two 2N×N or N×2N PUs or four N×N PUs. Since then, some investigations have been made to the prediction block structure based on HM1.0, including the removal of N×N prediction partition mode for all coding units (CU) except the smallest CU, and the removal of 4×4 Inter prediction. Experimental results show that with both these simplifications, the encoder complexity can be greatly reduced at the minimum cost of coding efficiency. Therefore both of these two suggestions were adopted by the HEVC standard.