Construction method of brain corpus callosum segmentation prediction image for corpus callosum state evaluation
A technology for predicting images and constructing methods, which is applied in the field of medical image segmentation and deep learning, and can solve problems such as high error rate, low detection rate of fetal corpus callosum abnormalities, and inability to accurately calculate the volume of corpus callosum
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0036] Embodiment 1 The construction of the fetal ultrasound image state analysis deep neural network model of the present invention
[0037] (1) Image preprocessing
[0038] a. Acquisition of ultrasound images of the fetal brain and segmentation label images of the corpus callosum
[0039] Ultrasound images of the fetal brain were collected by the TRT33 frequency-variable dual-plane brain probe using brightness modulation type ultrasonic section imager; the segmentation label image of the corpus callosum of the brain was provided by medical imaging technicians according to the fetal brain ultrasound image annotation;
[0040] b. Image data preprocessing
[0041]The acquired ultrasound images of the fetal brain are translated, transformed, twisted and enhanced, and elastically deformed, and the corner point detection and center point detection are performed on the corpus callosum segmentation label image of the brain. The specific method is as follows:
[0042] ① Use the hor...
Embodiment 2
[0069] Embodiment 2 Fetal ultrasound image state analysis of the present invention
[0070] Take a fetal brain ultrasound image data to be evaluated and input it into the deep neural network model constructed in Example 1, and construct a brain corpus callosum segmentation map through the output initial contour line and active contour offset to evaluate the state of the fetal brain corpus callosum , for the specific frame structure of corpus callosum state analysis based on deep neural network fetal ultrasound images, see image 3 .
[0071] In summary, the present invention converts the segmentation of the corpus callosum into the initial contour establishment and active contour convergence, uses the encoding and decoding module to obtain multi-scale image feature information, predicts the corpus callosum state code and initial contour of the fetal ultrasound image, and distributes the vector through the key points The construction of the sum loss function makes the key poin...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com