3D Dental Model Registration via Arch Curve Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for registering digital 3D models of teeth, such as ICP, often fail to align models that are severely misaligned or obtained at different times, limiting their use in diagnostic and analytical applications.
Innovation Solution
A transform method that computes a smooth arch curve using non-linear manifold embedding techniques, incrementally aligns the models using ICP, and selects the best initial starting point based on distance scores to achieve robust registration, even for full-arch and quadrant scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If ICP matching is used to register 3D models, then alignment refinement is achieved, but the method fails when models are severely misaligned or far apart in position or orientation
Solution Approach 1:
The system performs preliminary actions by computing arch curves and performing coarse alignment before applying ICP refinement. This preliminary positioning brings severely misaligned models into a state where ICP can effectively refine the alignment, resolving the contradiction between achieving precision and maintaining reliability across all alignment scenarios
Solution Approach 2:
The patent introduces intermediary elements (arch curves and coarse alignment transformations) that mediate between severely misaligned models and the ICP algorithm. These intermediaries bridge the gap by first establishing a rough correspondence, allowing ICP to then refine the alignment without failing due to initial severe misalignment
2Manufacturing precision
If good initialization is provided for ICP, then alignment refinement works effectively, but models obtained at different times or severely misaligned cannot be adequately registered
Solution Approach 1:
The system performs preliminary actions by computing arch curves and performing coarse alignment before applying ICP refinement. This preliminary positioning brings severely misaligned models into a state where ICP can effectively refine the alignment, resolving the contradiction between achieving precision and maintaining reliability across all alignment scenarios
Solution Approach 2:
The patent changes parameters by introducing arch curve-based coarse alignment as a preliminary step that transforms the initial parameter space. This allows the system to handle models from different time points by first establishing a rough correspondence through arch curve matching, then refining with ICP, thereby expanding adaptability while maintaining precision
3Loss of information
If synchronized display of multiple 3D models is implemented, then diagnostic comparison capability is enhanced, but computational complexity and processing time increase
Solution Approach 1:
The system segments the registration process into distinct phases: arch curve computation, coarse alignment, and ICP refinement. This segmentation allows each phase to handle specific aspects of the registration task independently, reducing overall system complexity while maintaining the ability to preserve diagnostic information through multi-model synchronized display
4Reliability
If arch curve computation and coarse alignment are performed before ICP, then severely misaligned models can be registered, but processing time increases
Solution Approach 1:
The system performs preliminary actions by computing arch curves and performing coarse alignment before applying ICP refinement. This preliminary positioning brings severely misaligned models into a state where ICP can effectively refine the alignment, resolving the contradiction between achieving precision and maintaining reliability across all alignment scenarios
Solution Approach 2:
The patent applies skipping by rapidly performing the preliminary arch curve computation and coarse alignment steps to quickly bring models into a suitable state for ICP refinement. This rushing through of preliminary steps minimizes the time penalty while ensuring registration success, as the subsequent ICP refinement is much more efficient when starting from a pre-aligned state
Data Source
Figure 1~2
Figure 3A~3B
Figure 4A~4B
AI summary
Systems and methods for displaying synchronized views and animations of digital 3D models of a person's intra-oral structure such as teeth. Two digital 3D models obtained from scans at different times are displayed in side-by-side views and synchronized via registration of the two scans or corresponding models. A user's control input to one displayed model causes the same manipulation of both models since they are registered. The two digital 3D models can also be displayed in an animation mode where the first model slowly morphs into the second model to illustrate changes in the intra-oral structure over time.