摘要
针对压缩传输对立体视频降质损伤的影响,进行大量主观实验获取评价数据,设计了一套合理的数据处理方法,并对处理后的实验结果进行了分析探究。该处理方法主要包括淘汰异常者、剔除异常值和均值化处理3个步骤。实验结果表明,所述方法处理数据的能力明显优于传统的格鲁布斯检验法。利用上述实验数据处理方法,分析经H.264压缩算法处理的立体视频实验数据,得出结论:量化参数(QP)的大小影响立体视频质量损伤度大小;参考视点携带的信息更占据主导地位;高分辨率视频有助于降质补偿;女性接纳视觉损伤的能力高于男性。探究所得结论为立体视频通信技术的优化提供技术支持。
Aiming at the influence of compression and transmission on the degradation damage of stereo videos, a reasonable data processing method is designed based on the evaluation data obtained from lots of subjective evaluation experiments, and then the final processed results are analyzed. The method includes three aspects： weeding out abnormal cases, removing abnormal values and averaging values. The experimental results show that this method is obviously better than the traditional Grubbs test. The data of stereo videos processed by the H.264 compression algorithm are analyzed by the method above, and the conclusions are drawn： the quantization parameter （QP） affects the damage degree of video quality; information carried by the reference viewpoint takes a more dominant place; high-resolution videos contribute to the compensation of degradation; the ability of females to accept visual impairment is stronger than that of males. The conclusions provide technical support for the communication technology optimization of stereo videos.
出处
《激光与光电子学进展》
CSCD
北大核心
2014年第9期113-121,共9页
Laser & Optoelectronics Progress
基金
国家自然科学基金(61002028)
国家自然科学基金重点资助项目(60932007)
国家863计划(2012AA011505
2012AA03A301)
关键词
成像系统
立体成像
主观评价
降质损伤
量化参数
数据处理
imaging systems
stereo imaging
subjective assessment
compression degradation
quantizationparameter
data processing
