Department of Physical Education, Dalian University of Foreign Languages, Dalian, Liaoning 116044, China.
School of Physical Education, Liaoning Normal University, Dalian, Liaoning 116029, China.
Comput Intell Neurosci. 2021 Dec 27;2021:9615290. doi: 10.1155/2021/9615290. eCollection 2021.
The quality of boxing video is affected by many factors. For example, it needs to be compressed and encoded before transmission. In the process of transmission, it will encounter network conditions such as packet loss and jitter, which will affect the video quality. Combined with the proposed nine characteristic parameters affecting video quality, this paper proposes an architecture of video quality evaluation system. Aiming at the compression damage and transmission damage of leisure sports video, a video quality evaluation algorithm based on BP neural network (BPNN) is proposed. A specific Wushu video quality evaluation algorithm system is implemented. The system takes the result of feature engineering of 9 feature parameters of boxing video as the input and the subjective quality score of video as the training output. The mapping relationship is established by BPNN algorithm, and the objective evaluation quality of boxing video is finally obtained. The results show that using the neural network analysis model, the characteristic parameters of compression damage and transmission damage used in this paper can get better evaluation results. Compared with the comparison algorithm, the accuracy of the video quality evaluation method proposed in this paper has been greatly improved. The subjective characteristics of users are evaluated quantitatively and added to the objective video quality evaluation model in this paper, so as to make the video evaluation more accurate and closer to users.
拳击视频的质量受到许多因素的影响。例如，在传输之前需要进行压缩和编码。在传输过程中，会遇到丢包和抖动等网络条件，这会影响视频质量。结合提出的影响视频质量的九个特征参数，本文提出了一种视频质量评估系统的架构。针对休闲体育视频的压缩损伤和传输损伤，提出了一种基于 BP 神经网络（BPNN）的视频质量评估算法。实现了特定的武术视频质量评估算法系统。该系统将 9 个拳击视频特征参数的特征工程结果作为输入，视频主观质量分数作为训练输出。通过 BPNN 算法建立映射关系，最终得到拳击视频的客观评价质量。结果表明，使用神经网络分析模型，本文中使用的压缩损伤和传输损伤特征参数可以获得更好的评估结果。与比较算法相比，本文提出的视频质量评估方法的准确性有了很大提高。本文将用户的主观特征定量评估并添加到客观视频质量评估模型中，从而使视频评估更加准确，更接近用户。
