Medical Image Region Extraction Using Group-Specific Learned Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face detection methods, such as those described in JP2007-128127A and JP2009-070099A, are ineffective in accurately extracting target regions from medical images.
Innovation Solution
An information processing apparatus and method that extracts information indicating a subject's physique from medical images, classifies the subject into groups based on age, gender, or species, and generates learned models using machine learning to accurately identify regions within medical images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single learned model is used for all subjects, then the device complexity is reduced, but the measurement precision of target region extraction deteriorates
Solution Approach 1:
The patent segments the learning task by creating separate learned models for different subject groups (e.g., adults and children). The specification unit classifies subjects into groups based on physique information, and the generation unit creates group-specific learned models. This segmentation allows each model to be optimized for its target group, improving extraction accuracy without requiring an overly complex universal model.
Solution Approach 2:
The patent applies local quality by making the learned model adaptive to different subject groups. Instead of a uniform model for all subjects, the system generates specialized models with local optimizations for specific groups (adults, children, etc.). This allows each model to have quality tailored to its intended use case, improving overall measurement precision while managing complexity through targeted specialization.
2Measurement precision
If group-specific learned models are generated, then the target region extraction accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-classifying subjects into groups using the specification unit before performing target region extraction. The physique information extraction and group specification are performed in advance, allowing the system to select or generate the appropriate learned model beforehand. This preliminary classification simplifies the overall process by avoiding the need for complex real-time model selection or switching during extraction.
Solution Approach 2:
The specification unit acts as an intermediary between the input image and the learned models. It extracts physique information and determines the appropriate subject group, thereby mediating between the general input and the specific group-oriented models. This intermediary layer simplifies model management by providing a clear classification mechanism that maps subjects to appropriate models without requiring complex coordination between multiple specialized models.
Data Source
AI summary
Provided are an information processing apparatus, an information processing method, and a program capable of accurately extracting a target region from a medical image. An information processing apparatus includes an extraction unit that extracts information indicating a physique of a subject from an image acquired by imaging the subject, a specification unit that specifies a group into which the subject is classified by using the information indicating the physique of the subject extracted by the extraction unit, and a generation unit that generates a learned model for each group through machine learning using, as learning data, image data indicating a medical image acquired by imaging the subject for each group and information indicating a region extracted from the medical image.