Video stream encoding quality evaluation method based on motion characteristics

Abstract

The invention discloses a video stream encoding quality evaluation method based on motion characteristics. The motion characteristics of videos are combined with the average perceived quality of a video set determined by a code rate, and the slope of an average quality curve of the videos is adjusted by the motion characteristics to achieve video stream encoding quality evaluation. The method comprises the following steps: step 1, establishing an average quality evaluation model of the video set, comprising: step 1.1, acquiring subjective perceived quality scores of video sequences, step 1.2, establishing an average perceived quality evaluation model of the video set based on the code rate; and step 2, establishing a video stream encoding quality evaluation model based on the motion characteristics, comprising: step 2.1, defining the motion characteristics of the video sequences, defining the motion characteristics of the video sequences from three angles, which are separately a motion vector size, a motion sector direction and a motion block ratio, step 2.2, acquiring a relationship between the motion characteristics and video time complexity characteristics, and step 2.3, establishing the video stream encoding quality evaluation model based on the motion characteristics of the videos.

Description

technical field [0001] The invention belongs to the field of multimedia communication, and in particular relates to a video stream encoding quality evaluation method based on motion characteristics. Background technique [0002] The end use of video and images is for people to watch. Video quality evaluation aims to obtain the human eye's perception of video image quality through some methods, including subjective evaluation methods and objective evaluation methods. The subjective evaluation is carried out under strict experimental environment and experimental process, asking a group of observers to rate the video quality. The objective evaluation method intends to measure the video quality by simulating human eyes through computing models or evaluation equipment. Subjective evaluation is the most accurate way to obtain a quality score. However, due to the strict requirements of the experiment, time-consuming, labor-intensive, and not suitable for real-time monitoring occ...

AI Technical Summary

Problems solved by technology

However, this paper is aimed at low bit rate and low resolution variable frame rate video in mobile applications (the image resolution is in SIF format), so the model is not suitable for standard definition or even high definition video, and the model does not analyze each video in detail Relationship between motion characteristics and subjective perceived quality

Images