Objective video quality assessment method for evaluating effects of freeze distortion in arbitrary video scenes

With the development of the broadband IP networks, video distribution, streaming, and communication services over networks are rapidly becoming common. To provide these services appropriately, we must design, monitor, and manage their quality based on subjective video quality. Therefore, we need an objective quality assessment method that enables video quality to be evaluated from video characteristics easily and quickly. We have already proposed an objective video quality assessment method for freeze distortion in network video services by considering the perceptual characteristics of short freeze distortion. In this previous method, we derived the objective video quality under the assumption that all freeze lengths in the video are constant. However, it was impossible to derive objective video quality when arbitrary freeze lengths occur because of network degradation in real services. We propose an objective video quality assessment method for arbitrary freeze lengths by extending the previous method and confirm that it can estimate subjective video quality with good accuracy for real applications. The correlation coefficient between subjective and objective quality is 0.94.