Endoscopic Image Mapping Using Anatomical Landmarks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image display devices for endoscopic examinations lack the ability to accurately evaluate omission in imaging, as they do not provide clear differentiation between imaged and non-imaged regions on anatomical outlines, making it difficult to determine if essential imaging areas have been captured.
Innovation Solution
An image display device and method that uses a computer to generate a map image by assigning endoscopic images to a virtual model based on connection relationships set using imaging order, feature patterns, and imaging conditions, allowing for accurate representation of imaged and non-imaged regions, and displaying notifications for any overlapping essential imaging regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a mosaic image is connected based on imaging position and outline shape for convenience, then the display is easier to understand, but the accuracy of evaluating imaging omission deteriorates
Solution Approach 1:
The patent introduces an anatomical landmark image as an intermediary reference that bridges the mosaic image and the outline view. By detecting anatomical landmarks (such as blood vessel patterns or tissue structures) in the mosaic image and matching them with corresponding landmarks on the outline view, the system achieves accurate spatial correspondence without relying merely on convenience-based registration. This intermediary landmark matching mechanism enables precise evaluation of imaging omission while maintaining clear visual understanding.
2Reliability
If essential imaging parts are set according to examination type, then the evaluation of imaging completeness becomes more meaningful, but the ability to accurately determine if essential parts are imaged deteriorates without proper registration
Solution Approach 1:
The patent replaces the manual or convenience-based registration mechanism with an automated image processing and pattern recognition system. By using computer-based detection of anatomical landmarks and automated matching algorithms, the system precisely determines the correspondence between mosaic images and outline views. This substitution of automated image processing for manual registration enables accurate evaluation of whether essential imaging parts have been captured, making the evaluation both meaningful and precisely measurable.
3Productivity
If rough registration is performed based on imaging position, then the processing is simpler and faster, but the position accuracy of mosaic image portions on the outline view deteriorates
Solution Approach 1:
The patent performs preliminary rough registration based on imaging position to quickly establish an initial correspondence, then applies subsequent refinement using anatomical landmark detection and matching. This two-stage approach allows the system to first achieve fast processing through rough alignment, then improve position accuracy by detecting anatomical landmarks (such as characteristic blood vessel patterns or tissue structures) in the mosaic image and matching them with the outline view. This preliminary action followed by refinement maintains processing efficiency while achieving high position accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There are provided an image display device, an image display method, and a program capable of accurately evaluating omission in imaging. A computer (22) that executes an image display method according to a program includes: a storage unit (53) that stores an endoscopic image of an imaging target organ; a connection relationship setting unit (60) that sets a connection relationship between a plurality of endoscopic images (S4); a landmark image detection unit (61) that detects a landmark image including an anatomical landmark corresponding to the imaging target organ among the plurality of endoscopic images (S6); a mapping unit (62) that assigns the landmark image to a landmark portion of a virtual model corresponding to the imaging target organ (S8) and assigns the plurality of endoscopic images to corresponding portions of the virtual model using the connection relationship set with the landmark image as a base point (S10); a map image generation unit (63) that generates a map image, which shows an imaged region and a non-imaged region of the imaging target organ, based on the virtual model in which the plurality of endoscopic images are assigned to respective portions (S12); and a display control unit (64) that displays the map image on a monitor (S13).