医学图像处理装置、医学图像处理方法、以及存储介质
By using convolutional neural networks and interactive methods in medical image processing devices, multiple candidate regions are automatically set and the best region of interest is selected, which solves the problem of inaccurate region of interest setting and improves the accuracy and reliability of quantitative analysis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CANON KK
- Filing Date
- 2021-10-08
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, methods for setting regions of interest are difficult to guarantee in terms of accuracy and consistency, which affects the accuracy of quantitative analysis, especially in medical image analysis such as bone mineral density and breast density.
Using a medical image processing device, multiple candidate regions are automatically or interactively set using methods such as convolutional neural networks and thresholding. The most suitable region of interest is determined through morphological processing and user selection, and quantitative analysis is performed.
It improves the accuracy and consistency of setting regions of interest, enhances the accuracy and reliability of quantitative analysis, provides a variety of visualization and selection methods for candidate regions, and supports users in selecting the best region based on patient information and image features.
Smart Images
Figure CN116456904B_ABST