沈丽丽, 吴子阳. 立体视频延时的事件相关电位研究[J]. 工程科学学报, 2018, 40(4): 508-515. DOI: 10.13374/j.issn2095-9389.2018.04.015
引用本文: 沈丽丽, 吴子阳. 立体视频延时的事件相关电位研究[J]. 工程科学学报, 2018, 40(4): 508-515. DOI: 10.13374/j.issn2095-9389.2018.04.015
SHEN Li-li, WU Zi-yang. Study of asynchronous stereoscopic video based on event related potentials[J]. Chinese Journal of Engineering, 2018, 40(4): 508-515. DOI: 10.13374/j.issn2095-9389.2018.04.015
Citation: SHEN Li-li, WU Zi-yang. Study of asynchronous stereoscopic video based on event related potentials[J]. Chinese Journal of Engineering, 2018, 40(4): 508-515. DOI: 10.13374/j.issn2095-9389.2018.04.015

立体视频延时的事件相关电位研究

Study of asynchronous stereoscopic video based on event related potentials

  • 摘要: 立体视频在无质量保证的互联网上传输时会引发立体视频左右视图的帧延时,影响观看质量.本文提出一种利用主观评价与客观脑电(electroencephalogram,EEG)相结合的方式分别对深度和水平延时运动立体视频刺激进行分析,并对二者的差异性进行比较.共有十名被试参加实验,被试观看随机呈现的不同延时等级包括无延时、延时1帧、延时2帧和延时3帧的立体视频片段,并对是否感知到延时效应做出主观判断,同时记录被试的EEG信号.从EEG信号中提取出事件相关电位(event related potentials,ERPs)并结合主观行为数据进行分析与比较.实验结果表明,为保证观看质量,深度运动立体视频中所能存在的最大延时帧数为1帧,水平运动中则不能存在延时效应.与水平延时运动相比,深度延时运动刺激产生的P300成分幅值变化范围更大,表明在深度延时运动刺激下大脑的活跃程度更高.同时,相同延时帧数的深度运动与水平运动刺激所产生的P300成分中,深度延时运动的潜伏期更长,这表明处理深度延时运动刺激所需时间更长.

     

    Abstract: When stereoscopic videos are transmitted on the no-quality-assurance internet, frame delay occurs in the left and right views, which affect the stereoscopic sensation as well as the viewing quality. In this paper, the stimulus of motion-in-depth and motionin-lateral asynchronous stereoscopic videos were analyzed and they were compared via qualitative and quantitative assessments. In this quantitative experiments, the well-known electroencephalogram (EEG) was utilized. A total of ten subjects was recruited to participate in an experiment, in which they watched randomly presented asynchronous stereoscopic video clips. For the qualitative assessment, after viewing one clip, the subjects were asked whether they perceived any asynchronicity in the frames. The degrees of asynchronicity included no delay, one frame delay, two frame delay, and three frame delay. The electroencephalogram (EEG) signal of each subject was recorded as they viewed the clip. Next, the event-related potentials (ERPs) were extracted from the EEG signals and the behavioral data and the ERPs were analyzed to compare the different reactions to the depths and lateral extents of the asynchronous stereoscopic motion. The results reveal that the maximum allowable asynchronicity perceived for motion-in-depth stereoscopic video clips is one frame, whereas there must be no delay in the motion-in-lateral stereoscopic video clips. In addition, compared with being stimulated by the motion-in-lateral asynchronous stereoscopic video clips, a magnitude of P300 has a wider range when the subjects are stimulated by motion-in-depth asynchronous stereoscopic video clips. This indicates that the brain becomes more active in this circumstance. At the same time, the latency of P300 is longer than when stimulated by the motion-in-lateral asynchronous stereoscopic video clips, which means that when the subjects are stimulated by the same degree of asynchronicity of motion-in-depth and -in-lateral stereoscopic video clips, more time is required to process the stimulation of motion-in-depth stereoscopic video clips.

     

/

返回文章
返回