Video Quality Assessment Using Block Analysis and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video conferencing, the increasing demand for real-time video services over limited bandwidth leads to network performance degradation, such as packet delay, jitter, and loss, resulting in poor quality of experience (QoE) due to channel congestion and the inability to efficiently manage resources in mobile environments.
Innovation Solution
An image quality assessment apparatus and method using block analysis and a machine learning algorithm to assess video stream quality, forming to-be-assessed blocks and inputting them into a quality assessment model for determining image quality, which can improve the encoding mechanism based on subjective user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video bandwidth is increased to improve video quality, then image quality is improved, but network bandwidth consumption increases and becomes unsustainable in mobile environments
Solution Approach 1:
The patent applies local quality by dividing the video image into multiple blocks and selectively encoding different blocks with different quality levels. Important regions (e.g., faces, objects of interest) are encoded with higher quality while less important regions use lower quality encoding. This allows the system to maintain perceived video quality while reducing overall bandwidth consumption, directly resolving the contradiction between image quality and bandwidth usage.
2Reliability
If network resources are increased to maintain video quality, then quality of service is improved, but cost and deployment complexity increase
Solution Approach 1:
The patent implements self-service through the quality assessment model that automatically evaluates video quality and identifies important regions without human intervention. The system autonomously determines which blocks require higher quality encoding based on content analysis, eliminating the need for manual resource allocation and reducing deployment complexity while maintaining reliable quality of service.
3Productivity
If machine learning algorithms are used to improve video coding, then coding efficiency is improved, but computational complexity increases
Solution Approach 1:
The patent reduces computational complexity through segmentation by dividing the video into multiple blocks and processing each block independently through the quality assessment model. This parallelizable approach allows the machine learning algorithm to run efficiently on distributed or mobile devices, maintaining improved coding efficiency while managing computational complexity through divided processing tasks.
Data Source
AI summary
An image quality assessment apparatus and an image quality assessment method are provided. In the method, multiple to-be-assessed blocks are formed for an image in a video stream, these to-be-assessed blocks are inputted to a quality assessment model, and a quality of the image is determined according to an output result of the quality assessment model. The quality assessment model is trained based on a machine learning algorithm.