Video loss prediction model in wireless networks

Abstract

This work discusses video communications over wireless networks (IEEE 802.11ac standard). The videos are in three different resolutions: 720p, 1080p, and 2160p. It is essential to study the performance of these media in access technologies to enhance the current coding and communications techniques. This study sets out a video quality prediction model that includes the different resolutions that are based on wireless network terms and conditions, an approach that has not previously been adopted in the literature. The model involves obtaining Service and Experience Quality Metrics, such as PSNR (Peak Signal-to-Noise Ratio) and packet loss. This article outlines a methodology and mathematical model for video quality loss in the wireless network from simulated data and its accuracy is ensured through the use of performance metrics (RMSE and Standard Deviation). The methodology is based on two mathematical functions, (logarithmic and exponential), and their parameters are defined by linear regression. The model obtained RMSE values and standard deviation of 2.32 dB and 2.2 dB for the predicted values, respectively. The results should lead to a CODEC (Coder-Decoder) improvement and contribute to a better wireless networks design.

Video Quality Loss Model on Communication Networks: An Approach Based on Frame Loss

“…Finally, the results are listed in Table V, confirming the approach's accuracy. Furthermore, the RMSE values obtained for all videos and resolutions were lower than a similar study found in [13]. Therefore, it is proven that adopting the video visual The designed model performance has proven more accurate for videos with less movement, e.g., "Ducks Take Off" and "Into Tree", which have lower RMSE values at each resolution.…”

“…The videos are encoded with H264/AVC and have a GOP (Group of Picture) size of 25 and a bitrate of 32 Mbps [13]. Each video has 500 frames, with a duration of 10 seconds.…”

“…Following an analysis carried out previously, a linear regression based on the Least Squares Method (Equation 4) technique is employed to predict PSNR loss (Equation 3) in terms of the video sequence in terms of percentage of Overall Frame Loss (OFL). For that, the PSNR metric with a logarithmic base function is adopted to evaluate video quality, as seen in [13]. The OF L values will be used in matrix A, and the L (Equation 5) values will be used in Matrix B to calculate the α and β coefficients (coef ) of Equation 4.…”

“…In papers [3], [9], [13], [14], authors have discussed the correlation between QoS and QoE, checking how robust the video compression standards are in error-prone communication systems. While in [15], authors investigate how to reduce the bandwidth consumption required by the streaming service.…”

Comparing simulation model for objective QoE video evaluation with real IPTV test scenario during appearance of packet losses

“…In addition, video service delivery in real time has been tested in [26] and [27]. Simulations of video service testing models are presented in [28] and [29]. This paper represents an extended version of paper published on conference TELFOR 2019.…”