Bronchus Image Segmentation via Inhalation Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical image analysis is hindered by the difficulty in segmenting bronchus images taken during exhalation due to collapse, making it challenging to analyze changes in bronchus dimensions compared to inhalation images where segmentation is easier.
Innovation Solution
An image processing apparatus and method that registers and transforms images taken at different times to generate segment images, allowing for comparison of bronchus changes in thickness, cross-sectional area, diameter, and ratios between inhalation and exhalation images, enhancing segmentation performance during exhalation by using transformation information from inhalation images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image segmentation is performed directly on exhalation images, then the processing is straightforward, but the segmentation accuracy is poor due to bronchus collapse
Solution Approach 1:
The patent performs image registration and transformation beforehand to generate a synthetic exhalation image from inhalation images before segmentation. This preliminary action creates optimal segmentation conditions by combining the anatomical detail from inhalation images with the collapsed bronchus structure from exhalation images, thereby improving segmentation accuracy without directly segmenting the difficult exhalation images
Solution Approach 2:
The patent introduces a synthetic exhalation image as an intermediary between the original inhalation images and the final segmentation result. This intermediary image is generated by registering and transforming inhalation images to match exhalation anatomy, serving as a bridge that provides both structural accuracy and anatomical detail for improved segmentation
2Measurement precision
If multiple images are processed to improve segmentation, then segmentation accuracy improves, but processing time increases
Solution Approach 1:
The patent extracts only the essential transformation information (registration parameters, scaling factors) from the image matching process, rather than processing entire image datasets. By extracting and storing only the critical transformation data, the system reduces processing time for subsequent segmentation operations while maintaining high segmentation accuracy
Solution Approach 2:
The patent performs image registration and transformation calculations in advance before segmentation is needed. This preliminary processing allows the system to store pre-computed transformation information that can be quickly applied during segmentation, reducing the time required at the actual segmentation stage
Data Source
AI summary
An image processing apparatus is disclosed. The image processing apparatus of the present invention comprises: an image receiving unit for receiving a first image and a second image of the same object taken at different times; a processor for obtaining transformation information by registering the first image on the basis of the second image, obtaining a first segment image corresponding to an area of the object from the first image, and generating a second segment image corresponding to an area of the object of the second image by transforming the obtained first segment image according to the transformation information; and an output unit for outputting the second segment image.