Image quality evaluation method and device, readable storage medium and terminal equipment
By calculating the feature distance between the generated image and the image generated reference information on multiple evaluation dimensions and aggregating it, the problem of difficulty in accurately evaluating the generated image quality of the generative artificial intelligence in the prior art is solved, and a comprehensive and accurate evaluation of the image quality is achieved.
CN119648610APending Publication Date: 2025-03-18SHENZHEN INTELLIFUSION TECHNOLOGIES CO LTD +1
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202411114567.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2025-03-18
AI Technical Summary
Technical Problem
The prior art is difficult to accurately evaluate the image quality generated by generative artificial intelligence.
Method used
By calculating the feature distances between the generated image and the image generation reference information on multiple evaluation dimensions (such as image painting style, shooting angle, content expression), and aggregating these feature distances, and using a preset image quality evaluation model for evaluation.
Benefits of technology
A comprehensive and accurate evaluation of the generated image quality is achieved, and feature information can be comprehensively considered from multiple dimensions to improve the accuracy and comprehensiveness of the evaluation.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention belongs to the technical field of image processing, and particularly relates to an image quality evaluation method and device, a computer readable storage medium and terminal equipment. The method comprises the following steps: acquiring a generated image; calculating the feature distance between the generated image and the image generation reference information in each evaluation dimension; wherein the image generation reference information comprises a cue word for generating the generated image and a target image expected to be generated, and the feature distance is used for representing the difference degree of the generated image and the image generation reference information in the evaluation dimension; aggregating the feature distances to obtain an aggregated feature distance; performing image quality evaluation on the generated image by using a preset image quality evaluation model based on the aggregated feature distance; wherein the image quality evaluation model is an artificial intelligence model which is obtained through pre-training and is used for carrying out image quality evaluation.
Need to check novelty before this filing date? Find Prior Art