Attribute-Based Image Filtering for Medical Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical image diagnosis, it is challenging to efficiently designate specific tomographic images from a large set of three-dimensional medical images captured by imaging apparatuses like CT and MRI, as the existing methods do not effectively filter or prioritize images based on attribute information, leading to difficulty in focusing on relevant images.
Innovation Solution
An information processing apparatus and method that assigns independent attribute information to images, allowing only images with designated attribute information to be displayed, using a slider bar with limited operable range based on this information, enabling easy selection of images of interest, such as regions of interest or lesions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the number of tomographic images in a three-dimensional medical image is increased, then the image coverage and diagnostic capability are improved, but it becomes difficult to designate and focus on specific target images
Solution Approach 1:
The patent segments the large set of tomographic images into multiple groups based on attribute information (e.g., anatomical region, lesion type, imaging parameters). Each group represents a subset of images sharing common characteristics, making it easier for users to navigate and designate target images within manageable segments rather than searching through all images individually
Solution Approach 2:
The patent introduces attribute information as an intermediary layer between the user and the images. Instead of directly browsing images, users interact with attribute tags that filter and organize images. This intermediary system enables efficient image designation by translating user requirements into targeted image selection based on matching attributes
2Loss of information
If all tomographic images are made displayable, then complete image information is available, but the user cannot easily focus on images of interest
Solution Approach 1:
The patent applies preliminary action by pre-assigning attribute information to each tomographic image before display. This advance organization allows the system to quickly filter and present only relevant images when a user specifies particular attributes, eliminating the need to display all images while ensuring complete information is available on demand
Solution Approach 2:
The patent implements local quality by making image displayability conditional on specific attribute matches. Instead of uniform display of all images, the system selectively displays images that possess the locally relevant attributes needed for the current diagnostic task, thereby focusing user attention on pertinent images while maintaining access to the full image set
3Ease of operation
If a slider bar is provided for selecting images, then image selection capability is improved, but the operable range becomes too large when many images are present
Solution Approach 1:
The patent segments the continuous slider bar operable range into multiple discrete sections, each corresponding to a specific group of images with shared attributes. This segmentation allows the slider to navigate through organized image subsets rather than spanning the entire image set continuously, reducing the effective operable range while maintaining comprehensive selection capability
Solution Approach 2:
The patent changes the parameter of slider bar operation from direct position mapping to attribute-based filtering. Instead of moving the slider to select images based on position alone, the system interprets slider movements in conjunction with attribute filters, dynamically adjusting the operable range to reflect only those images matching current attribute criteria, thereby reducing the effective search space
Data Source
AI summary
An information processing apparatus including at least one processor, wherein the processor is configured to: acquire a plurality of images to which mutually independent attribute information is assigned; and limit the images, among the plurality of images, to images that are displayable on a display based on the attribute information.