Lesion Boundary Correction in 3D Medical Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately detecting lesions in 3-dimensional images is challenging due to noise, low resolution, and low contrast, especially when lesion boundaries are complex and blurred, making it difficult to correct lesions across multiple 2-dimensional images.
Innovation Solution
An apparatus and method that select a candidate image frame by generating lesion values from 2-dimensional images, extracting candidate frames based on these values, and using a reference image frame to correct lesions in other frames, allowing for accurate and efficient lesion detection and correction across a 3-dimensional image set.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a lesion is detected in a 3-dimensional image consisting of multiple 2-dimensional images, then the lesion detection coverage is improved, but the correction accuracy deteriorates due to the complexity of correcting lesions across multiple frames
Solution Approach 1:
The patent segments the 3-dimensional image into multiple 2-dimensional image frames, allowing independent lesion detection in each frame while maintaining overall 3D coverage. The correction process then operates frame-by-frame using reference frames, breaking down the complex multi-frame correction problem into manageable individual frame corrections that can be propagated systematically.
Solution Approach 2:
The patent introduces reference image frames as intermediaries to facilitate accurate lesion correction across multiple frames. These reference frames serve as mediators that contain accurate lesion information and are used to correct lesions in other frames through image segmentation and boundary extraction, enabling precise correction without directly comparing all frames against each other.
2Loss of information
If multiple initial contours are defined at opposite edges of an object for multiple slice contour detection, then the contour detection completeness is improved, but the device complexity increases
Solution Approach 1:
The patent extracts only the necessary reference frames containing accurate lesion information from the entire 3-dimensional image set, rather than processing all frames equally. This extraction approach maintains complete contour detection by selecting frames that capture all lesions, while reducing complexity by focusing computational resources only on these extracted reference frames and their corresponding correction targets.
3Adaptability or versatility
If lesion correction is performed on each 2-dimensional image frame independently, then the correction flexibility is improved, but the overall correction accuracy deteriorates due to inconsistency across frames
Solution Approach 1:
The patent implements a feedback mechanism where corrected lesions from reference frames are used to guide and verify corrections in other frames. The system continuously refines lesion boundaries by comparing corrections across multiple frames and using accurate reference corrections as feedback to adjust and validate corrections in non-reference frames, ensuring consistency while maintaining flexibility.
Data Source
Figure 1
Figure 2(a)~2(g)
Figure 3(a)~3(g)
AI summary
An apparatus for extracting a candidate image frame includes a generating unit configured to generate at least one lesion value that represents a characteristic of a lesion included in each of a plurality of 2-dimensional image frames that form a 3-dimensional image, and an extracting unit configured to extract, from the image frames, at least one candidate image frame usable for correcting a boundary of the lesion based on the at least one lesion value.