Method and device for performing flow control on multimedia data and method and device for training flow control model

Abstract

The invention relates to a method and a device for performing flow control on multimedia data and a method and a device for training a flow control model. The method comprises the following steps: performing combination processing on communication state data and current coded data in a current period to obtain to-be-processed combined data; performing self-attention mechanism processing on the to-be-processed combined data to obtain intermediate processing features, performing multi-task classification based on the intermediate processing features, and outputting target flow control data of at least one dimension; and based on target coded data determined by the target flow control data, triggering flow control processing on multimedia data generated in the next period in the multimedia communication process. Due to the fact that different types of coding data and communication state data can be subjected to importance degree distinguishing through a self-attention mechanism, the primary and secondary derivation function of high-level characterization features is fully played, and therefore when the output target flow control data is used for flow control processing of the next period, the output target flow control data can better adapt to the complex and changeable real network environment, and the flow control effect is good.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, in particular to a method and device for performing flow control on multimedia data and training a flow control model. Background technique [0002] With the development of network technology, more and more multimedia data transmission scenarios require real-time monitoring of network status, so as to realize flow control of multimedia data transmission according to the monitored network status. Taking the multimedia data transmission scenario as an example of a voice or video real-time call scenario, the voice or video real-time call is usually implemented by using a network RTC (Real-Time Communication, real-time call). In an RTC call, it is usually necessary to monitor the network status in real time, and then modify the overall call configuration according to the real-time network status. For example, if the network status is good, the encoding rate can be increased...

AI Technical Summary

Problems solved by technology

[0003] In the traditional solution, the GCC algorithm in WebRtc is widely used in the current adaptive bit rate control algorithm, which is the network congestion control algorithm for real-time media communication. However, when the GCC algorithm is actually used, there will be a certain delay, and Excessive reliance on empirical configuration leads to inaccurate flow control in actual control

Images