No-reference adaptive streaming media quality evaluation method and system optimized by neural network

Abstract

The invention provides a neural network-optimized non-reference adaptive streaming media quality evaluation method, which calculates the freeze perception index through the freeze information; calculates the video quality switching damage and the video oscillation damage through the video quality change characteristics, and combines the video compression quality to form Combining video quality switching and video oscillation damage video integration quality; calculating video video quality features, and using video quality features, freeze times, total freeze duration and video duration as input features to establish a neural network model and perform fitting training; Integrate the freeze perception index, video integration quality and fitting results to obtain the evaluation results of the end user experience quality. At the same time, a system for executing the above method is provided. The present invention requires few relevant parameters of the freeze buffering event, improves the adaptability and feasibility in practical application, and introduces the consideration related to the time effect through the neural network to increase the comprehensiveness and accuracy of the evaluation.

Description

technical field [0001] The present invention relates to a no-reference quality evaluation technology in the technical field of adaptive streaming media services, in particular to a neural network-optimized no-reference adaptive streaming media quality evaluation method and system. Background technique [0002] With the deep integration of network technology and multimedia technology, streaming video services relying on the Internet have gradually become an important part of people's daily life and work. At the same time, diversified and portable terminal devices continue to be popularized and promoted. More and more users watch videos online on mobile devices such as tablet computers or smart phones, which makes streaming video services develop rapidly. [0003] Traditional streaming media distribution technologies can generally be divided into two categories: one is real-time streaming transmission technology, which mainly includes Real Time Transfer Protocol (Real Time Tra...

AI Technical Summary

Problems solved by technology

But at the same time, there are certain disadvantages in real-time streaming transmission: the transmission process needs to match the bandwidth of the network connection. When problems such as network congestion occur, data transmission errors or loss may occur, resulting in poor video quality.

[0009] 1. Due to the existence of the Primacy Effect and the Recency Effect, the current evaluation method of the freeze effect needs to collect the time and duration of each freeze during video playback, and there is currently a lack of lightweight The parameter collection tool available for deployment;

[0010] 2. The relationship between some key features and quality of experience is uncertain, such as video quality switching frequency, recency effect and primacy effect, and the impact of different quality levels on perceived quality. Different research results have emerged in different research processes. In conclusion, it is difficult to explicitly map the relationship between the two in a deterministic way

[0011] At present, there is no description or report of the similar technology of the present invention, and no similar data at home and abroad have been collected yet.

Images