Image quality detection method and pet image processing method

By using an image quality detection model to perform global quality assessment of medical images, this technology solves the problem that the region of interest assessment in existing technologies cannot accurately predict the overall image quality, thus achieving efficient and accurate image quality detection.

CN122245655APending Publication Date: 2026-06-19SHANGHAI UNITED IMAGING HEALTHCARE +1
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHANGHAI UNITED IMAGING HEALTHCARE
Filing Date
2025-10-24
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

In medical image quality detection, existing technologies cannot accurately predict the overall image quality by assessing the image quality of the region of interest, resulting in poor accuracy of the detection results.

⚗Method used

An image quality detection model is used to perform global quality assessment of medical images. A convolutional neural network is used to score the quality of each pixel and generate a quality assessment result image, which directly evaluates the quality of the entire image.

🎯Benefits of technology

It eliminates the need for segmentation or selection of regions of interest, simplifies the operation process, reduces the impact of uncertainty, accurately predicts the quality of the entire image, and improves the precision and accuracy of the detection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This application relates to an image quality detection method and a PET image processing method, belonging to the field of image processing technology. When detecting the quality of medical images, it can efficiently obtain accurate and reliable image quality detection results. The method includes: acquiring a medical image to be detected; inputting the medical image into a trained image quality detection model, whereby the image quality detection model determines the image quality information of the medical image, the image quality information including multiple quality quantification indicators, each quality quantification indicator representing the quality score of at least one pixel; the image quality detection model is trained based on a sample medical image and a corresponding label image, the label image including the image quality information corresponding to the sample medical image; and determining the image quality detection result of the medical image based on the image quality information corresponding to each pixel in the medical image.
Need to check novelty before this filing date? Find Prior Art