A medical image data processing system based on deep learning

CN122244076APending Publication Date: 2026-06-19EAST CHINA UNIV OF SCI & TECH +1
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
EAST CHINA UNIV OF SCI & TECH
Filing Date
2026-03-24
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing medical image processing systems struggle to simultaneously meet the requirements of comprehensive suppression of large areas of useless background and accurate extraction of small lesion areas when segmenting MRI image regions. Furthermore, image noise and artifacts can easily cause image region boundary drift, leading to inaccurate segmentation.

Method used

MRI images are divided into direct and indirect images, and different segmentation strategies are adopted: direct images are segmented by spreading outwards, while indirect images are segmented by converging inwards. Muscle region features are extracted using a deep learning model, and trigger conditions are set for ordered feature extraction.

Benefits of technology

It achieves accurate segmentation and feature extraction of image regions, improves the efficiency of medical image data processing, and ensures adaptive segmentation and automatic feature recognition of MRI images of different qualities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122244076A_ABST
    Figure CN122244076A_ABST
Patent Text Reader

Abstract

This invention relates to the field of medical data processing technology, and discloses a medical image data processing system based on deep learning. It includes an image segmentation module for dividing MRI images into direct and indirect images; a region aggregation module for generating image regions from the direct and indirect images; a feature extraction module for extracting muscle region features; and a feature analysis module for generating analysis results. This invention can accurately segment smaller image regions from larger MRI images, avoiding the negative interference of useless background regions in MRI images on subsequent feature extraction. It can also perform adaptive segmentation operations on MRI images of different qualities, ensuring that direct and indirect images maintain an orderly and independent segmentation effect, achieving sequential segmentation of direct and indirect images, and avoiding the problem of inaccurate image region segmentation caused by single image region segmentation operations.
Need to check novelty before this filing date? Find Prior Art