3D Image Segmentation Using Selective 2D/3D Geodesic Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image segmentation methods require manual refinement of each 2D slice in a 3D image stack, which is time-consuming and prone to inconsistent results, while lacking efficient 3D computation strategies.
Innovation Solution
A geodesic distance algorithm that combines 2D and 3D calculations, allowing for user-driven 3D image segmentation by locking interacted slices and performing 3D computations only between these slices, reducing the need for full 3D computations on every modification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual refinement of each 2D slice is performed, then segmentation accuracy can be improved, but time consumption increases significantly
Solution Approach 1:
The patent segments the 3D image processing into discrete 2D slice operations. By treating each 2D slice independently and using 3D geodesic distance computation only where necessary (between non-contiguous slices), the method divides the complex 3D segmentation problem into manageable 2D tasks, improving both efficiency and accuracy.
Solution Approach 2:
The patent applies 3D geodesic distance computation selectively rather than uniformly across all slices. By performing full 3D computation only between non-contiguous slices with user interactions and using 2D computation for intermediate contiguous slices, the method applies partial action to reduce computational time while maintaining segmentation accuracy where it matters most.
2Stability of the object's composition
If full 3D computation is performed on every modification, then segmentation consistency is improved, but computational time increases
Solution Approach 1:
The patent implements local quality by differentiating computation strategies based on slice position and user interaction. Non-contiguous slices with user modifications receive full 3D geodesic distance computation for consistency, while intermediate contiguous slices use efficient 2D computation, optimizing the balance between consistency and efficiency locally at each slice.
Solution Approach 2:
The patent introduces dynamic computation adaptation where the level of computation (2D vs 3D) changes based on user interactions and slice contiguity. The system dynamically adjusts computational intensity, performing 3D computation only when necessary to maintain consistency while using 2D computation for routine updates, thereby improving productivity without sacrificing consistency.
3Productivity
If 3D geodesic distance computation is minimized, then computational time is reduced, but segmentation accuracy may deteriorate
Solution Approach 1:
The patent performs preliminary 2D geodesic distance computation on all slices before applying 3D computation. This preliminary action establishes a baseline segmentation that is then refined with 3D computation only where needed, ensuring accuracy is maintained while minimizing unnecessary 3D computations and improving overall computational speed.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A computer-implemented method comprising: receiving a 3D image including an object depicted in the image, the 3D image comprising an ordered set of 2D images; determining a contour around the object in a first of said 2D images; and determining a contour around the object in a second of said 2D images, the second 2D image being non-contiguous with the first in said ordered set, having an intermediate region comprising one or more intermediate ones of said 2D images between the first and second 2D images within said ordered set. In each of the first and second 2D images, inside of the contour is classified as foreground and outside of the contour is classified as background. The method further comprises performing a 3D geodesic distance computation to classify points in the intermediate region as foreground of background.