Fast transcoding from H.264 to HEVC based on region feature analysis

The new video coding standard, High Efficiency Video Coding (HEVC), achieves much higher coding efficiency than the state-of-the-art H.264. Transcoding H.264 video to HEVC video is important to enable gradual migration to HEVC. Therefore, a fast H.264 to HEVC transcoding algorithm based on region feature analysis is proposed. First, each frame is segmented into three regions in units of coding tree unit (CTU) based on the correlation between image coding complexities and coding bits of the H.264 source stream. Then the searching depth range of each CTU is adaptively decided according to the region type. After that, motion vectors are de-noise filtered and clustered in order to analyze the region features of coding unit (CU). Based on the analysis results, the minimum searching depth of CU and partitions of prediction unit (PU) are optimally selected, and the motion vector predictor and search window size of motion estimation are also optimally decided for further reduction of the computational complexity. Experimental results show that the proposed algorithm achieves a significant improvement on transcoding speed, while maintaining high Rate-Distortion performance.     

HEVC视频编码器快速算法

高效视频编码(HEVC)采用了多种先进的编码工具来大幅度提高压缩效率,但同时也带来了高的编码复杂度。考虑到LCU编码时需遍历所有的CU模式、PU模式和TU模式,来确定当前LCU的最优模式组合,其过程耗时较大。针对这点本文分别对CU、PU和TU提出模式判决过程优化算法,同时,每一种PU模式在进行运动矢量搜索时复杂度较高,继而提出运动估计优化方法。实验结果表明,相比于原始的HEVC编码方法,高速模式判决算法可将编码速度提高近15～20倍。     

基于背景帧的监控和会议视频帧间快速算法

为降低高性能视频编码（High Efficiency Video Coding,HEVC）中监控和会议视频的编码复杂度,减少编码时间,提出了一种基于虚拟背景帧的监控和会议视频帧间编码的快速编码单元（Coding Unit,CU）划分和预测单元（Prediction Unit,PU）模式选择算法。对原始视频序列进行像素统计分析生成虚拟背景帧,利用生成的虚拟背景帧对CU分类。通过分析不同类别的CU继续划分的比例和其纹理复杂度,快速决策当前CU的划分和PU的模式类型。这种利用虚拟背景帧做判决的算法通过减少CU深度和PU模式遍历而降低编码复杂度,达到快速帧间编码的效果。实验结果表明,与HM16.9相比,该算法在峰值信噪比（Peak Signal-to-Noise Ratio,PSNR）平均下降0.07 dB和码率平均增加1.93%的情况下,能平均减少33.31%的编码时间。     

Fast CU size decision and mode decision algorithm for intra prediction in HEVC

In the intra prediction process, High Efficiency Video Coding (HEVC) provides a quadtree-based coding unit (CU) block partitioning structure and up to 35 kinds of prediction modes to improve the coding performance. These technologies improve the coding efficiency significantly while the coding complexity is simultaneously increased rapidly as well. In this paper, a novel fast CU size decision and mode decision algorithm is proposed for the intra prediction of HEVC. The overall algorithm consists of two processes, the fast CU size decision and fast mode decision. In the fast CU size decision process, we adopt an adaptive discretization total variation (DTV) threshold-based CU size determination algorithm to skip some specific depth levels. In the fast mode decision process, an orientation gradient-based mode decision is proposed to reduce the candidate modes involved in the rough mode decision (RMD) and the rate distortion optimization (RDO) process. The experimental results on the HEVC reference software HM demonstrate that the proposed algorithm can significantly reduce the coding time with negligible performance loss.     

An early split and skip algorithm for fast intra CU selection in HEVC

The rapid development and increase of multimedia applications, as well as the demand for higher video-quality services at restricted resources such as storage capacity, transmission bandwidth and power consumption, has brought the urgent need for more efficient video compression techniques. The new video coding standard high efficiency video coding (HEVC) has a significant superiority over its predecessor advanced video coding (AVC). HEVC is reported to halve the bit rate with the same visual quality, or a better quality with the same bit rate when compared with AVC. Beside other improvements, HEVC significantly gets its power from the use of dynamic hierarchical quad-tree structure by partitioning the frames into smaller regions called coding units (CU), by means of a rate–distortion optimization process. However, this improvement yields to a dramatic increase of high computational complexity and increased encoding time, which primarily restricts its adaptation in real-time applications. In this paper, we proposed an early CU determination algorithm for fast encoder realization to reduce the encoding time which is the most important part of the standard standing for development. The experimental results show that the proposed algorithm has approximately 45 % encoding time saving with a 4.6 % bit-rate increment, on average.     

Fast HEVC I-frame coding based on strength of dominant direction of CUs

High efficiency video coding (HEVC), the latest international video coding standard, greatly outperforms previous standards such as H.264/AVC in terms of coding bitrate and video quality. The coding efficiency improvement in HEVC is achieved by introducing several new techniques such as recursive quad-tree structure and increased number of intra prediction modes. However, computational load is also increased due to employing the new techniques. In this paper, we propose a solution for fast I-frame coding in HEVC standard using homogeneity of Coding Units (CUs). The proposed solution consists of two stages. In the first stage, we evaluate CU homogeneity by computing a parameter named dominant direction strength and predict CU size by this means. In the second stage, we select 11 modes out of 35 for the specified CU size based on dominant direction of the CU. Experimental results indicate that the proposed method achieves on average 45.8 % reduction on coding time, with very similar coding efficiency as the HEVC reference software. Moreover, we designed tree-stage pipelined architecture for our method which can operate at 235 MHz maximum clock rate which means it can be used for real-time coding of all intra configuration of HEVC videos up to level 6.2.     

Motion and disparity vectors early determination for texture video in 3D-HEVC

3D-HEVC is the state-of-the-art video coding standard for 3D video, and it is an extension of high efficiency video coding (HEVC) standard. Besides the original HEVC coding tools, 3D-HEVC adopts some advanced coding tools, such as disparity vector (DV), inter-view prediction and inter-component prediction. However, these advanced tools lead to extremely high encoding complexity at the same time, thus it cannot be well applied in real-time multimedia systems. In this paper, we propose a motion and disparity vectors early determination algorithm to reduce 3D-HEVC computational complexity. First, based on the statistical analyses, the spatial and temporal motion vector (MV) candidates are adaptively reduced for the prediction unit (PU) with the Merge mode. Then, for the PU with the Inter mode, the combination of spatial and temporal candidates is used to early determine the final MV. Finally, an adaptive optimization algorithm is adopted to select the valid inter-view disparity vectors (DV) candidates. Moreover, if the difference between candidate vectors is within a conditional range, current PU will be encoded with the Merge mode to skip unnecessary coding process. Experimental results show that for the texture views encoding, the proposed algorithm achieves an average of 33.03% encoding time saving, and an average of 0.47% BD-Rate increases.

     

利用深度学习的HEVC帧内编码单元快速划分算法

新一代视频编码标准高效视频编码(High Efficiency Video Coding,HEVC)中编码单元(Coding Unit,CU)大小不同的特性使得编码效率得到显著提升,但同时带来了极高的计算复杂度.为了去除CU划分中多余的计算从而降低编码复杂度,本文提出了一种利用深度学习的编码单元快速划分算法.首先使用原始视频亮度块及编码信息建立了一个HEVC中CU划分的数据库,用于接下来本文深度学习神经网络的训练.然后,为了更好地贴合编码单元划分的层级结构,本文提出了一种基于Inception模块的神经网络结构,使之内嵌于HEVC编码框架中对编码单元的划分进行提前预测,有效地去除了All Intra配置下中冗余的CU划分计算.实验结果表明,本文提出的算法与HEVC官方测试模型(HM16.12)相比,编码时间平均降低了61.31%,而BD-BR与BD-PSNR仅为1.86%和-0.13dB.     

利用时空相关性的HEVC帧内编码块快速划分

目的 为了提升高效视频编码（HEVC）的编码效率,使之满足高分辨率、高帧率视频实时编码传输的需求。由分析可知帧内编码单元（CU）的划分对HEVC的编码效率有决定性的影响,通过提高HEVC的CU划分效率,可以大大提升HEVC编码的实时性。方法 通过对视频数据分析发现,视频数据具有较强的时间、空间相关性,帧内CU的划分结果也同样具有较强的时间和空间相关性,可以利用前一帧以及当前帧CU的划分结果进行预判以提升帧内CU划分的效率。据此,本文给出一种帧内CU快速划分算法,先根据视频相邻帧数据的时间相关性和帧内数据空间相关性初步确定当前编码块的编码树单元（CTU）形状,再利用前一帧同位CTU平均深度、当前帧已编码CTU深度以及对应的率失真代价值决定当前编码块CTU的最终形状。算法每间隔指定帧数设置一刷新帧,该帧采用HM16.7模型标准CU划分以避免快速CU划分算法带来的误差累积影响。结果 利用本文算法对不同分辨率、不同帧率的视频进行测试,与HEVC的参考模型HM16.7相比,本文算法在视频编码质量基本不变,视频码率稍有增加的情况下平均可以节省约40%的编码时间,且高分辨率高帧率的视频码率增加幅度普遍小于低分辨率低帧率的视频码率。结论 本文算法在HEVC的框架内,利用视频数据的时间和空间相关性,通过优化帧内CU划分方法,对提升HEVC编码,特别是提高高分辨率高帧率视频HEVC编码的实时性具有重要作用。     

基于决策树的HEVC编码单元划分算法

为了有效降低高效视频编码标准HEVC编码过程中的计算复杂度,提出了一种基于决策树的编码单元划分算法。该算法将编码单元划分问题归为分类问题,提取编码单元的划分信息到决策树模型中进行学习,得到决策树分类器。利用分类器对满足分类条件的编码单元进行划分,跳过了率失真计算,从而降低了视频编码的计算复杂度。实验结果表明,本算法在保证视频质量的同时有效降低了编码计算复杂度。     

基于运动特性的HEVC帧间模式快速决策算法

针对高性能视频编码采用四叉树结构大大增加了编码复杂度的问题,提出了一种基于运动特性的帧间模式快速决策算法。首先,对不同运动区域下的编码单元（Coding Unit,CU）块,利用当前CU与空时域相邻CU深度相关性减少当前CU深度的遍历范围;然后,依据当前CU与其时空域相邻CU及上一深度CU对应的预测单元（Prediction Unit,PU）在空间划分上的相似性,减少PU模式的遍历范围,加速帧间预测过程。实验结果表明,相比于HM16.9,在不同编码接入方式下该算法可平均降低54%左右的编码时间,且输出比特率增加较少。     

Square-type-first inter-CU tree search algorithm for acceleration of HEVC encoder

In this paper, a fast inter-coding algorithm is proposed to reduce the computational load of HEVC encoders. The HEVC reference model (HM) employs the recursive depth-first-search (DFS) of the quad-tree search in terms of rate-distortion optimization in selecting the best coding modes for the best CU, PU, TU partitions, and many associated coding modes. The proposed algorithm evaluates the RD costs of the current CU only for its square-type PUs in the top-down search of the DFS. When the CU partition with the square-type PU is better than its sub-level CU partitions in terms of RD cost in bottom-up search of the DFS, the square type of current CU partition, along with its coding mode, is selected as the best partition. Otherwise, non-square-type PUs for the current CU level are evaluated. If the sub-partition is better than the CU with the non-square PUs, the sub-partition is finally selected as the optimum PU. Otherwise, the best non-square PU is selected as the best PU for the current level. Experimental results demonstrate that the proposed square-type-first inter-PU search can reduce the computational load in average encoding time by 66.7 % with 1–2 % BD loss over HM reference software. In addition, the proposed algorithm can yield an additional average time saving of 26.8–35.5 % against the three fast encoding algorithms adopted in HM.     

HEVC快速编码深度选择算法

高效视频编码（High Efficiency Video Coding,HEVC）作为下一代新的视频编码标准,旨在有限网络带宽下传输高质量的网络视频。与现有的视频编码标准相比,高效视频编码具有更高的灵活性和压缩率。编码单元（Coding Unit,CU）是视频编码处理的基本单元,原有的算法通过四叉树递归获取最佳CU深度,在提高视频压缩性能的同时引入了较高的计算复杂度。针对该问题,提出了一种快速编码深度选择算法,该算法利用相邻CU的深度信息计算一个深度预测特征值,通过该特征值进行深度选择,以避免不必要的计算,降低计算复杂度。实验结果表明,该算法在保证视频压缩效果的同时有效降低了计算复杂度。     

基于纹理特性与空域相关的高效视频编码帧内分级快速算法

为了降低高效视频编码(HEVC)标准的编码复杂度,提出一种基于纹理特性与空域相关性的帧内分级快速算法。首先,采用最大编码单元(LCU)级的快速算法,通过利用相邻LCU的编码深度值加权预测得到当前LCU的预测深度,并利用块标准差和自适应阈值策略确定当前LCU的纹理复杂度,将当前LCU的预测深度和纹理复杂度相结合来预测当前LCU的最有可能深度范围(MPDR);其次,采用编码单元(CU)级的深度判决快速算法(CUDD-FA),将基于边缘图的CU深度预判策略和基于率失真(RD)代价相关性的CU提前中止策略相结合,实现了CU级深度的提前确定,进一步降低了帧内编码复杂度。与原始HEVC算法相比,所提算法在全I帧编码模式下编码时间平均减少41.81%,BD-rate(Bjøntegaard Delta bit rate)仅上升0.74%,BDPSNR(Bjøntegaard Delta Peak Signal-to-Noise Rate)仅降低0.038 dB;与代表性文献算法相比,所提算法在编码时间节省更多的情况下率失真性能更好。实验结果表明,在率失真性能损失可以忽略不计的前提下,所提算法能有效降低HEVC帧内编码复杂度,特别是高分辨率视频序列,有利于HEVC的实时视频应用。     

基于图形信息的HEVC帧间预测快速算法

由于最新的视频编码标准HEVC（high efficiency video coding）应用四叉树的递归结构进行编码单元的划分,使得帧间预测的过程极为复杂,编码的时间效率比较低下。针对HEVC帧间预测过程,提出了一种基于纹理信息和稳定区域检测的快速算法,旨在检测出相对平坦的区域和随时间变化较稳定的区域,来提前终止四叉树的编码单元的划分。这些平坦或者稳定的区域用较大的编码预测单元就能进行比较准确的运动估计和预测,不需要再进行更小编码预测单元的划分,因此HEVC编码器的时间复杂度大幅度降低。实验结果表明,改进后的算法编码时间平均节省了34%左右,并且造成的视频质量损失非常低,有0.038 dB的PSNR下降和0.56%的码率增加。     

高效视频编码帧内快速深度决策算法

针对新一代高效视频编码（HEVC）帧内预测中编码单元（CU）的编码深度选择过程中计算复杂度较高的问题,提出了一种基于空域相关性的帧内快速深度决策算法。首先,利用相邻已编码树单元（CTU）的深度通过线性加权得到当前CTU深度估计值;然后,对当前CTU深度估计值设置较为合适的深度双阈值提前终止编码树单元的划分或跳过CTU的某些深度,来缩小当前CTU的深度范围,从而减少不必要的深度计算。实验结果表明：与HM12.0相比,所提算法对比较简单的视频序列编码时间的减少比较明显,在亮度峰值信噪比（Y-PSNR）几乎不变的情况下（平均降低0.02 dB）,编码时间平均减少了34.6%。此外,所提算法容易与其他算法进行融合,能进一步降低HEVC的帧内计算复杂度,最终达到实时传送高清视频的目的。     

高效率视频编码快速帧内预测算法

针对高效率视频编码(HEVC)帧内预测过程中,编码单元四叉树划分算法计算复杂度极高的问题,提出一种基于多重纹理特征的HEVC帧内编码单元快速划分算法,该算法能够缩小划分的深度区间。首先,使用自定义的纹理提取方法提取出编码单元中的多重纹理特征;其次,使用支持向量机(SVM)对多重纹理特征参数进行训练,得出决策函数;最后,根据决策函数,跳过前面不必要的划分和提前终止划分。实验结果表明,同原始HM 12.0相比,快速划分算法编码时间平均减少43.23%,码率平均增加0.84%,明显提高了帧内编码效率。此外,所提算法容易与其他算法进行融合,进一步降低HEVC的帧内计算复杂度。     

针对HEVC编码单元的二分深度划分算法

为了降低高效率视频编码（HEVC）的编码单元（CU）进行四叉树递归遍历的时间,提出一种改进的编码单元快速划分算法.首先,利用帧间时间域的相关性,提取前一帧相同位置CU的最优划分结构,以预测当前CU的划分深度;然后通过改进编码CU结构划分遍历的算法,减少CTU （Coding Tree Unit）四叉树结构的遍历,即从二分深度开始遍历,在每一步遍历之前,判断是否提前终止遍历.实验表明,与HM15.0中的基准划分算法相比,本文算法能够在保证编码性能的同时,降低了55.4%的编码时间,提高了HEVC的编码效率.     

Fast coding unit (CU) determination algorithm for high-efficiency video coding (HEVC) in smart surveillance application

High-efficiency video coding (HEVC) is a successor to the H.264/AVC standard as the newest video-coding standard using a quad-tree structure with the three block types of a coding unit (CU), a prediction unit (PU), and a transform unit (TU). This has become popular to apply to smart surveillance systems, because very high-quality image is needed to analyze and extract more precise features. On standard, the HEVC encoder uses all possible depth levels for determination of the lowest rate-distortion (RD) cost block. The HEVC encoder is more complex than the H.264/AVC standard. An efficient CU determination algorithm is proposed using spatial and temporal information in which 13 neighboring coding tree units (CTUs) are defined. Four CTUs are temporally located in the current CTU and the other nine neighboring CTUs are spatially situated in the current CTU. Based on the analysis of conditional probability values for SKIP and Merge modes, an optimal threshold value was determined for judging SKIP or Merge mode according to the CTU condition and an adaptive weighting factor. When SKIP or Merge modes were detected early, other mode searches were omitted. The proposed algorithm achieved approximately 35 % time saving with random-access configuration and 29 % time reduction with low-delay configuration while maintaining comparable rate-distortion performance, compared with HM 12.0 reference software.     

Boundary correlation-based intracoding for SHVC algorithm and its efficient VLSI architecture

Scalable high-efficiency video coding (SHVC) can provide variable video quality according to terminal devices. However, a computational complexity of SHVC is increased by introducing new techniques based on high-efficiency video coding (HEVC). In this paper, a hardware-oriented low complexity algorithm is proposed for the reference software of SHVC (SHM11.0). In our proposed algorithm, an optimal coding unit depth is determined by analyzing the boundary correlation in a coding tree unit before encoding starts. Simulation results show that the proposed algorithm can achieve over 62% computation complexity reduction comparing to the original SHM11.0. Compared with other related work, over 11% time saving has been achieved without PSNR loss. Moreover, to confirm the efficacy of the proposed algorithm, a hardware architecture is designed targeting on the CU depth decision algorithm. Synthesis results show that the hardware cost is about 1.8K gate and achieve a scalable working clock frequency in the case of FPGA (CycloneV) implementation.