A Novel No-Reference QoE Assessment Model for Frame Freezing of Mobile Video

In this paper, a novel no-reference (NR) Quality of Experience (QoE) assessment model for frame freezing of mobile video is proposed. Four source video sequences with smooth motion intensity which were extracted from LIVE mobile database have been used to create different types of test sequences. Two subjective experiments are conducted with these distorted sequences, and the Differential Mean Opinion Scores (DMOS) are obtained. Then a QoE model is proposed based on the experimental results. This model can quantitatively measure the perceptual quality of users’ experience when they are watching the frame freezing videos. Due to the lack of publicly available datasets, we establish a new database of mobile videos with frame freezing distortion based on the LIVE mobile database. The proposed model is compared with three other QoE assessment metrics on the new database, and the result shows the proposed model has a better performance than others.