• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>IEEE International Conference on Recent Trends In Electronics Information Communication Technology >Content based search mode improvement for B frame in HEVC
【24h】

Content based search mode improvement for B frame in HEVC

机译：HEVC中B帧基于内容的搜索模式改进

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Motion Estimation is most important part of almost all video coding standards such as H.262/MPEG-2, H.264/MPEG-4 and High Efficiency Video Coding (H.265/HEVC). The HEVC is latest video compression standard and has incorporated many improvements like bitrate reduction, increased coding efficiency and compression efficiency but at the cost of increased complexity. The complexity of H.265 is increased due to Motion Estimation (ME), as it consumes almost half time of total time. Motion Estimation algorithm used in reference software code is robust and it is independent of motion information of videos. Real world video sequences show a wide range of motion related information which varies from uniform to random. If the motion characteristics of video sequences are taken into account earlier, it is possible to develop a fast motion estimation algorithm for video sequences. The proposed algorithm includes motion vector prediction which uses motion data of neighboring candidates and motion classification using the characteristics of video sequences. Based on the type of motion the algorithm chooses between different search patterns to enhance the matching process. Early termination helps to additional speedup and improves the process. The proposed method is going to be evaluated using different metrics like percentage reduction in coding time, PSNR, Bit Rate and RD graph.
机译：运动估计是几乎所有视频编码标准（例如H.262 / MPEG-2，H.264 / MPEG-4和高效视频编码（H.265 / HEVC））中最重要的部分。 HEVC是最新的视频压缩标准，并结合了许多改进功能，例如降低比特率，提高编码效率和压缩效率，但代价是增加了复杂性。 H.265的复杂性由于运动估计（ME）而增加，因为它消耗了总时间的近一半时间。参考软件代码中使用的运动估计算法是鲁棒的，并且独立于视频的运动信息。现实世界中的视频序列显示了各种与运动有关的信息，这些信息从统一到随机变化。如果较早地考虑了视频序列的运动特性，则有可能为视频序列开发一种快速运动估计算法。所提出的算法包括运动矢量预测，该运动矢量预测使用相邻候选者的运动数据，并且利用视频序列的特征进行运动分类。基于运动的类型，算法在不同的搜索模式之间进行选择，以增强匹配过程。提前终止有助于进一步加快速度并改善流程。拟议的方法将使用不同的度量标准进行评估，例如减少编码时间百分比，PSNR，比特率和RD图。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号