Early TU Split Termination in HEVC Based on Quasi-Zero-Block

Abstract

Abstract-The new introduced concept, transform tree or residual quadtree (RQT) in HEVC standard, brings high coding performance along with high computational complexity because the optimal TU partition is estimated in a ratedistortion sense at the encoder sides by testing all kinds of partitions. This paper focus on early TU split termination in terms of the tradeoff between computation and coding quality. An early TU split termination scheme based on quasi-zeroblock (QZB) is proposed, which is defined by two aspects about the quantized transform coefficients; the sum of all absolute coefficients and the number of nonzero coefficients. Besides, the selective probability of every TU depth is calculated and analyzed. Experimental results show that the proposed method (HM-QZB) could achieve 22.8% reduction in encoding time and 50.59% reduction in TU processing time compared to the HEVC test model HM10.0 encoder with about 0.04dB BD-PSNR loss in coding performance.

Fast HEVC Encoding Decisions Using Data Mining

“…In [21] and [22], the computational complexity required to decide the RQT structure is reduced by early terminating the recursive TU splitting based on the number of non-zero transformed coefficients. However, as the computational complexity reductions achieved when constraining RQTs is very limited, not many works exploit simplifications of the procedures used to determine these structures.…”

“…Several authors have recently proposed heuristics for simplifying the partitioning decision for Coding Units (CU), Prediction Units (PU) and Transform Units (TU), aiming at decreasing the computational complexity of HEVC without hurting its compression efficiency [9][10][11][12][13][14][15][16][17][18][19][20][21][22]. Even though these works succeed in their goal of decreasing the encoding complexity up to a certain extent, most of them only deal with a single type of partitioning structures (either CUs [9][10][11][12][13], PUs [14][15][16][17] or TUs [21,22] separately), which limits the achievable complexity reductions.…”

“…Even though these works succeed in their goal of decreasing the encoding complexity up to a certain extent, most of them only deal with a single type of partitioning structures (either CUs [9][10][11][12][13], PUs [14][15][16][17] or TUs [21,22] separately), which limits the achievable complexity reductions. Besides, the majority of them result in non-negligible losses in R-D efficiency.…”

State of the Art on Computational Complexity of Video Encoders

“…In [56] and [57], the computational complexity required to decide the RQT structure is reduced by early terminating the recursive TU splitting based on the number of nonzero transformed coefficients. The method in [56] proposes the concept of quasi-zero-blocks (QZB), which are defined by criteria based on the values of two quantities: the sum of the absolute values of coefficients and the number of nonzero coefficients.…”

“…The method in [56] proposes the concept of quasi-zero-blocks (QZB), which are defined by criteria based on the values of two quantities: the sum of the absolute values of coefficients and the number of nonzero coefficients. According to the authors, subtle differences between nonzero blocks and blocks with very small coefficients cannot be perceived by human eyes, so that QZBs can be used in an early termination scheme to stop splitting of TUs.…”