Video Encoding Quality Assessment Using Content Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing objective video quality assessment methods lack accuracy in predicting encoding quality, as they do not adequately consider subjective human perception and are limited by incomplete models that only account for encoding information such as bit rate, codec, and network packet loss.
Innovation Solution
A method and device that assess video encoding quality by determining the quantization parameter and content complexity of a video stream, using acquired quantization parameters and bytes per pixel to predict encoding quality, which better reflects human subjective experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing objective video quality assessment methods are used, then the assessment process is simple, but the accuracy of predicting encoding quality is low
Solution Approach 1:
The video stream is divided into multiple slices, and quality assessment is performed on each slice separately using quantization parameters and byte-per-pixel metrics. This segmentation allows for more precise local quality measurement while maintaining computational efficiency through parallel processing of slice-level data.
Solution Approach 2:
The assessment method transitions from using only encoding information parameters (bit rate, codec) to incorporating content-specific parameters (quantization parameter, bytes per pixel, content complexity). This parameter expansion enables the model to capture both encoding quality and content characteristics, significantly improving prediction accuracy.
2Measurement precision
If existing models only consider encoding information, then the model is simple to implement, but it cannot accurately reflect subjective feelings of human eyes
Solution Approach 1:
The assessment model combines multiple types of information: encoding parameters (quantization parameter), content metrics (bytes per pixel), and derived features (content complexity). This composite approach integrates both technical encoding data and content-specific characteristics, creating a comprehensive quality assessment that better reflects human subjective perception.
Solution Approach 2:
The model incorporates content complexity calculations that provide feedback about the actual video content being assessed. By calculating bytes per pixel and deriving content complexity metrics, the system adapts its assessment to the specific content characteristics, enabling more accurate prediction of human visual perception across different video types.
Data Source
AI summary
A method and a device for assessing video encoding quality. The method includes: acquiring a quantization parameter of a slice of a video frame and a quantity of bytes per pixel of the slice of the video frame of the video stream; determining complexity of content of the video according to the quantity of bytes per pixel of the slice of the video frame of the video stream; and predicting the video encoding quality according to the complexity of content of the video and the quantization parameter of the video. In the present invention, the complexity of content of the video is also considered in predicting the video encoding quality. Therefore, encoding quality predicted by a model that is obtained by considering the complexity of content of the video better satisfies subjective feelings of human eyes, thereby improving accuracy of prediction.