Advances in Multimedia

Research Article

Nonintrusive Method Based on Neural Networks for Video Quality of Experience Assessment

Table 7

Papers with machine learning methods for assessing QoE.
Paper# video sequencesResolutionCode cSSIMVQMPSRNMSEBRFRPLR value (decodificables frames rate)QP Playout interruptionsDelayJitterBWMLBSGOP Machine learning Assess MOS/DMOS?Correlation performance
PCCSROCCRMSEOR
He et al. [17], 20121QCIFMPEG4  X   X   XX X Probabilistic Neural Network (PNN)Yes0.9286NoNoNo
Kipli et al. [18], 2012N/DNo dataNo dataX XX           BPNNYes0.891NoNoNo
Singh et al. [19], 2012 4HD-720pH.264      X XXX    RNNYesNoNo0.37No
Lia et al. [20], 20114No dataNo data               BPNN NoNoNoNoNo
Khan at al. [21], 20106QCIFH.264    XX       X ANFISYes 0.8717 No0.2812No
Du et al. [22], 20091SDNo data  X   X   XXXX BPNNYesNoNoNoNo
Piamrat et al. [23], 20091No dataH.264  X   X      X PSQA with RNNYesNoNoNoNo
Khan et al. [24], 20083CIFMPEG4    XXXX    X  ANFISYesR = 0.8056 for MOS predicted versus MOS Measure and R = 0.9229 for predicted versus measure NoRMSE = 0.1846 for MOS predicted versus MOS measure and RMSE = 0.06234 for predicted versus measure No
Rubino et al. [25], 2004Only VoIPNo dataNo data      X      X RNNNoNoNoNoNo
BR: bitrate, FR: frame rate, PLR: Packet Loss Rate, QP: Quantization Parameter, BW: bandwidth, MLBS: Mean Loss Burst Size, GOP: Group of Pictures, PCC: Pearson correlation coefficient, SROCC: Spearman rank order correlation coefficient, RMSE: Root Mean Square Error, and OR: Outlier Ratio.