Cerebrovascular Structure Standardization via Hierarchical Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for analyzing cerebrovascular images struggle to standardize and normalize vessel branches, making it difficult to diagnose and treat cerebrovascular diseases effectively.
Innovation Solution
A method involving an analysis device that receives cerebrovascular images, extracts vascular unit structures, calculates feature values, and classifies these structures using pretrained learning models to generate standardized cerebrovascular structure information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional methods are used to analyze cerebrovascular images, then the analysis process is simple, but the ability to standardize and normalize vessel branches is insufficient
Solution Approach 1:
The patent segments the cerebrovascular structure into multiple hierarchical levels: vascular unit structures (basic units), chunks (groups of vascular units), and vessel branches (assemblies of chunks). This segmentation enables standardized analysis by breaking down the complex vasculature into manageable, classifiable units that can be processed systematically through multiple learning models.
Solution Approach 2:
The patent employs pretrained learning models that have been trained in advance on large datasets of cerebrovascular images. These models perform preliminary classification of vascular structures into standardized categories before final analysis, enabling consistent normalization across different subjects and populations without requiring real-time complex processing.
2Measurement precision
If detailed classification of vascular structures is performed, then diagnostic accuracy is improved, but processing time increases
Solution Approach 1:
The classification process is divided into multiple stages with different levels of detail. First, vascular unit structures are classified into chunks using a pretrained first learning model. Then, chunks are classified into vessel branches using a pretrained second learning model. This hierarchical segmentation allows detailed classification to be performed efficiently by distributing the computational load across multiple simpler classification tasks rather than one complex task.
Solution Approach 2:
The learning models are pretrained in advance on large datasets, performing the computationally intensive learning work before actual analysis. During actual processing, the pretrained models quickly classify new vascular structures by applying learned patterns, significantly reducing processing time while maintaining high classification accuracy.
3Reliability
If population-based standardization is implemented, then diagnostic reference value is improved, but data processing complexity increases
Solution Approach 1:
The patent standardizes population data by segmenting it into the same hierarchical structure used for individual analysis: vascular unit structures, chunks, and vessel branches. This consistent segmentation allows systematic comparison between individual subjects and population norms, enabling reliable diagnostic references while maintaining manageable data processing through structured organization.
Solution Approach 2:
The same pretrained learning models and classification framework used for individual analysis are applied to population data standardization. This universal approach allows the system to handle both individual subject analysis and population-based reference creation using identical methods, simplifying the overall data processing architecture while improving diagnostic reliability through population comparison.
Data Source
AI summary
The method of generating standardized cerebrovascular structure information includes extracting a plurality of vascular unit structures from a cerebrovascular image of a subject, extracting feature values of each of the plurality of vascular unit structures, classifying the plurality of vascular unit structures into chunks by the feature values of each of the plurality of vascular unit structures, classifying multiple vessel branches composed of vascular unit structures belonging to the same chunk by feature values of each of the vascular unit structures, and dividing at least one of the multiple vessel branches into a predetermined number of segments and setting indices for all the segments or segments at certain intervals among the segments.