Image Quality Evaluation Network Feature Shift Mechanism
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image quality evaluation methods are inefficient and labor-intensive, relying on manual scoring which hampers accuracy and efficiency, especially when evaluating video frames.
Innovation Solution
An image quality evaluation method utilizing a pre-trained network that extracts features from images, performs shift operations to generate additional features, and combines them to determine image quality, improving efficiency and accuracy by automating the evaluation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual evaluation method is used, then evaluation process is simple, but evaluation efficiency is low
Solution Approach 1:
The patent replaces the manual mechanical evaluation process with an automated deep learning-based image quality evaluation network. The network automatically extracts features from video frames and determines quality metrics without human intervention, thereby dramatically improving evaluation efficiency while accepting increased system complexity through the use of neural network models.
2Measurement precision
If manual scoring is used, then evaluation process is straightforward, but evaluation accuracy is limited
Solution Approach 1:
The patent performs preliminary feature extraction from video frames before the actual quality evaluation. The image quality evaluation network pre-processes the video content by extracting relevant visual features, which are then used for accurate quality assessment. This preliminary action enables both high accuracy and efficient processing by preparing data in advance for the evaluation stage.
3Loss of information
If traditional feature extraction is used, then computational load is low, but feature comprehensiveness is insufficient
Solution Approach 1:
The patent segments the feature extraction process into multiple specialized components within the image quality evaluation network. Different neural network modules extract specific types of features (e.g., structural features, texture features, color features) separately, then combine them to form a comprehensive quality assessment. This segmentation enables thorough feature extraction while managing computational complexity through modular processing.
Data Source
AI summary
The present application discloses an image quality evaluation method and apparatus, a device, and a storage medium. The method includes: acquiring a to-be-evaluated image; and inputting the to-be-evaluated image into an image quality evaluation network to obtain an image quality evaluation result, where the image quality evaluation network is configured to extract a first image feature from an input image, perform a shift operation on the first image feature to acquire one or more second image features, and determine the image quality evaluation result by combining the first image feature and the acquired second image features; a size of each of the second image features is the same as a size of the first image feature, and regions with identical values of features exist in different positions between the first image feature and the second image features.