Cerebral Microbleed Detection via SWI and Phase Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting cerebral microbleeds using magnetic resonance images are inefficient due to the time-consuming nature of visual inspection and difficulty in distinguishing microbleeds from similar substances, even with automated CAD systems.
Innovation Solution
A device and method employing a preprocessing unit for normalizing and converting magnetic resonance images, combined with a YOLO neural network module for candidate region detection and a cerebral microbleeds determination neural network module for accurate identification, utilizing both sensitivity-weighted imaging and phase images to enhance detection performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual inspection of SWI images is used to detect cerebral microbleeds, then detection capability is achieved, but detection time is excessively long and detection accuracy is low
Solution Approach 1:
The patent replaces the manual visual inspection mechanism with an automated neural network-based image processing system. The CNN model automatically analyzes SWI images to detect cerebral microbleeds, eliminating the time-consuming manual review process while maintaining or improving detection accuracy through consistent algorithmic application.
Solution Approach 2:
The patent introduces an intermediate processing layer consisting of image preprocessing techniques and a convolutional neural network that acts as a mediator between the raw SWI image and the final detection result. This intermediary system enhances the visibility and detectability of microbleeds by applying normalization, contrast enhancement, and automated feature extraction before presentation to clinicians.
2Productivity
If automated CAD systems are used to improve detection efficiency, then processing speed increases, but detection accuracy remains insufficient due to difficulty in distinguishing microbleeds from similar substances
Solution Approach 1:
The patent employs parameter changes in the form of image preprocessing techniques that modify the visual parameters of SWI images. By applying normalization, contrast enhancement, and other preprocessing operations, the system optimizes the visual parameters to enhance the distinction between microbleeds and similar substances, thereby improving detection accuracy while maintaining automated processing speed.
Solution Approach 2:
The patent applies preliminary actions through extensive image preprocessing before the actual detection process. The CNN model receives preprocessed images that have already undergone normalization, contrast enhancement, and other optimizations, which prepare the data in advance to facilitate more accurate detection and reduce false positives during the automated analysis phase.
3Reliability
If manual inspection methods are used, then simple equipment is required, but detection reliability is low due to difficulty in distinguishing microbleeds from lime substances, iron and veins
Solution Approach 1:
The patent applies segmentation by dividing the detection task into distinct modular components: image preprocessing module, neural network detection module, and result interpretation module. This segmentation allows each component to be optimized independently, with the CNN model specifically trained to distinguish microbleeds from similar substances, thereby improving reliability while keeping the overall system manageable through modular architecture.
Solution Approach 2:
The patent replaces the manual inspection mechanism with an automated neural network system that consistently applies learned patterns for distinguishing microbleeds from similar substances. This substitution eliminates human variability and subjectivity, providing more reliable and consistent detection results across different cases and operators, despite the increased computational complexity.
Data Source
AI summary
A device and method for detecting cerebral microbleeds use magnetic resonance images. The disclosed device includes a preprocessing unit that normalizes an SWI image and a phase image, respectively, of the magnetic resonance images, and performs phase image conversion for inverting a code of the normalized phase image, a YOLO neural network module that receives a two-channel image in which the preprocessed SWI image and phase image are concatenated and detects a plurality of candidate regions for the cerebral microbleeds, and a cerebral microbleeds determination neural network module that receives patch images of candidate regions of the SWI image and phase image based on the candidate regions and determines whether the patch images of each candidate region are an image with a symptom of the cerebral microbleeds through a neural network operation.