Liver Segmentation in MR Images Using Multi-Channel Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual segmentation of the liver in magnetic resonance images is challenging due to variability in intensity units across scans, making global intensity cues unreliable for organ segmentation.
Innovation Solution
A method for fully automatic segmentation using a discriminative learning-based framework that utilizes multi-channel features from MRI images, employing statistical classifiers and a training database of liver shapes to establish relationships between features from multiple channels and the liver boundary, with Marginal Space Learning for pose estimation and fine-scale boundary localization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual delineation of the liver is performed, then segmentation can be achieved, but it is time-consuming and shows significant variability among users
Solution Approach 1:
The system performs automatic liver segmentation without human intervention by using training databases and statistical classifiers to autonomously identify and delineate liver boundaries in MR images, eliminating the need for manual user input while maintaining consistent segmentation quality
Solution Approach 2:
The patent replaces the manual mechanical process of user-based delineation with an automated computational system using training databases, feature extraction, and statistical classifiers to perform segmentation objectively and consistently
2Ease of operation
If global intensity cues are used for segmentation in MR images, then segmentation can be simplified, but intensity units vary across scans making them unreliable
Solution Approach 1:
The patent divides the segmentation process into multiple independent stages: initial boundary detection using training databases, pose estimation, and fine-scale boundary localization. Each stage uses specific features appropriate to that stage rather than relying on a single global intensity measure
Solution Approach 2:
The system transforms the segmentation approach by changing from relying on absolute intensity values to using relative intensity relationships and multi-channel feature combinations, which are invariant to the varying intensity units across different MR scans
3Device complexity
If single-channel MR images are used for segmentation, then processing is simpler, but segmentation accuracy is reduced
Solution Approach 1:
The patent combines information from multiple MR image channels by extracting features from each channel and integrating them through statistical classifiers, allowing the system to leverage the complementary information in each channel to improve overall segmentation accuracy
Data Source
AI summary
A method and system for fully automatic liver segmentation in a multi-channel magnetic resonance (MR) image is disclosed. An initial liver boundary in the multi-channel MR image, such as an MR Dixon scan. The segmented initial liver boundary in the multi-channel MR image is refined based on features extracted from multiple channels of the multi-channel MR image using a trained boundary detector. The features may be extracted from an opposed channel and a water channel of an MR Dixon scan.