An Objective Quality Assessment Method for Panoramic Video Based on Spatiotemporal Pipeline Modeling

Abstract

The invention discloses a panoramic video objective quality evaluation method based on space-time pipeline modeling, and the method comprises the steps: firstly adding a concept of time domain distortion according to a human eye visual mechanism, and building a video quality evaluation model based on a space-time pipeline; secondly, realizing the establishment of a space-time pipeline by a block search algorithm for motion estimation, the video content segmented spatially being associated front and back in time; the distortion of the space-time pipeline being decomposed into average distortionand time domain distortion, the average distortion being used for reasonably averaging the space distortion of blocks in the space-time pipeline through a recursive filter, and the recursive filter being used for consolidating the space-time pipeline distortion calculation by combining the frequency and amplitude of the distortion change gradient between the blocks; and finally, performing spatial domain pooling and time domain pooling calculation on all spatial-temporal pipeline distortions to obtain a final video quality score. By adopting the method provided by the invention, the accuracyof panoramic video objective quality evaluation can be greatly improved, so that the panoramic video objective quality evaluation better conforms to the subjective quality evaluation result of human eyes.

Description

technical field [0001] The invention belongs to the field of video coding and processing, and in particular relates to an objective quality evaluation method of panoramic video based on spatio-temporal pipeline modeling. Background technique [0002] With the continuous development of virtual and augmented reality technology, the formulation and improvement of technical standards for panoramic video media systems is becoming more and more important. Panoramic video (that is, 360-degree video) is to capture the image information of the entire scene through a professional camera, use software to splicing, and use a special player to play. When watching a 360-degree video, viewers usually need to wear a virtual helmet to watch video images in different positions through the movement of the head, so as to achieve the immersive and immersive sense of watching. However, since panoramic video contains a lot of viewing information, and usually has a large image resolution (such as ...

AI Technical Summary

Problems solved by technology

However, this method only considers the spatial characteristics of the panoramic video, and does not consider the time-domain transformation of distortion, and there is still room for improvement in accuracy.

Images