System for automatic bite registration from individual jaw geometries without physical registration

Abstract

Computer-implemented data processing system for the automatic determination of a relative positional relationship between the upper and lower jaw, in particular a maximum intercuspal position (MIP), comprising: > a) an input interface (110) configured to receive > at least two separately captured three-dimensional dental > geometries of an upper and a lower jaw; > > b) a pre-registration module (120) configured to determine > an initial spatial assignment of the individual jaw geometries > to each other based on geometric features of the dental arches; > > c) a fine-registration module (130) configured for iterative > refinement of the spatial assignment by minimizing a > distance or correspondence measure between corresponding > surface areas; > > d) a contact optimization module (140) configured to determine > an optimal relative position by maximizing the number or > quality of occlusal contact points while simultaneously minimizing > penetrations between the dental geometries;> > e\) an output interface (150) configured to provide > the specified relative positional relationship as a transformation matrix, > contact point list and / or registered model pair for downstream > analysis, simulation or manufacturing systems.;

Description

II. Technical Field

[0001] The invention relates to a data processing system and a registered dental bite position in the field of dental diagnostics and digital dentistry, in particular for the automatic determination of the maximum intercuspation position (MIP) from separately recorded three-dimensional single-jaw scans without a physical bite registration. III. State of the art

[0002] Determining the spatial relationship between the upper and lower jaw is a fundamental prerequisite for any dental functional analysis, prosthetic restoration, and orthodontic diagnostics. Conventionally, this relationship is established using a physical bite registration: The patient bites into a deformable material (wax, silicone, PVS) that fixes the relative position of the jaws to each other.

[0003] Existing registration methods have significant limitations: Physical bite registrations require patient contact and material costs (€5–30 per case), are susceptible to distortions due to material shrinkage, patient movement, or uneven biting, and do not generate directly machine-readable data. Transferring a physical registration to a digital system (e.g., scanner + articulator) introduces additional sources of error.

[0004] Modern intraoral scanners (Medit i700, iTero Element, 3Shape TRIOS, Shining3D Aoralscan) increasingly offer buccal scans, in which the patient bites down and an image of the posterior teeth in occlusion is captured. However, this method depends on the scanner quality, the opening angle, and the accessibility of the oral cavity, and does not provide reliable results in patients with partially edentulous teeth or limited mouth opening.

[0005] On the software side, manual alignment tools exist in CAD systems (exocad, 3Shape Dental System), which require the user to perform point-to-point mapping or manual positioning. Fully automatic registration algorithms for single-jaw dental scan pairs without a physical registration and without a buccal scan are not known.

[0006] Therefore, there is a need for a system that automatically determines the maximum intercuspation position from two separately recorded single-jaw scans—purely algorithmically, without physical registration, without buccal scan, and without manual assignment. IV. Objective of the invention

[0007] The invention is based on the objective of providing a data processing system that automatically determines the maximum intercuspation position (MIP) from two separately acquired three-dimensional single-jaw scans --- without physical bite registration, without buccal scan and without manual positioning by the user. V. Solution to the task

[0008] The task is solved by the data processing system according to claim 1, the registered dental bite position according to claim 10 and the arrangement according to claim 13. 1. Multi-stage registration workflow

[0009] The system operates in three stages: pre-registration (coarse), fine registration (iterative), and contact optimization (occlusion-specific). This three-stage architecture ensures that both global geometric features (arch shape) and local occlusal details (cusp-fossa relationships) are taken into account. 2. Pre-registration (Module 120)

[0010] The initial spatial alignment is derived from the dental arch geometry. In a preferred embodiment, the system analyzes the principal axes of both dental arches using PCA and determines an initial translation and rotation that brings the arches into an anatomically plausible relationship. The mandible is mirrored transversely and vertically approximated to the maxilla. This pre-registration does not need to be precise—it only needs to be close enough to the optimum so that the subsequent fine registration converges. 3. Fine registration (Module 130)

[0011] An iterative registration method refines the spatial mapping. In a preferred embodiment, a modified Iterative Closest-Point (ICP) method is used, which minimizes point-to-surface distances. The iteration typically converges after 20–50 steps. Alternative methods such as Coherent Point Drift (CPD) or probabilistic registration can be used. 4. Contact optimization (Module 140)

[0012] The key innovation: Following geometric fine-tuning, an occlusion-specific module optimizes the positional relationship with respect to maximum intercuspation. This involves calculating a multi-criteria evaluation function that maximizes occlusal contact points and minimizes penetrations. In a preferred embodiment, contacts are defined as surface points whose distance to the antagonist falls below a threshold value. Penetrations are defined as surface points that penetrate the antagonist.

[0013] The optimization module varies the relative position within clinically plausible limits (typically ±2 mm translation, ±3° rotation) and evaluates each candidate position. The position with maximum contact and minimum penetration is determined as the MIP. 5. Degrees of freedom constraint (Module 135)

[0014] The registration is limited to clinically relevant degrees of freedom. In a preferred embodiment, three degrees of freedom are primarily varied: vertical translation (bite height), anterior-posterior translation, and transverse translation. Rotations around the vertical axis are permitted within a limited range. Tilting around the transverse axis (mouth opening / protrusion) is fixed or severely restricted, as the MIP determination requires closed occlusion. 6. Confidence measure and validation (Module 160)

[0015] The system calculates a confidence level for the selected registration. In a preferred embodiment, this is based on the ratio of contact points to penetrations, the convergence rate, and the spatial distribution of contacts across the dental arch. A high confidence level indicates a reproducible, clinically plausible registration. A low confidence level warns of unreliable results and may indicate missing teeth, severe malocclusion, or insufficient scan data. 7. Browser-based implementation

[0016] A preferred embodiment is entirely browser-based (WebGL, in particular Three.js). The dental geometries are processed client-side and do not leave the user's computer (GDPR compliance). In a preferred embodiment, registration is completed in under 60 seconds on commercially available hardware. VI. List of reference symbols 10 OK-Mesh (upper jaw single scan) 11 UK-Mesh (lower jaw single scan) 20 Registered Bite Positions (Issue) 21 Contact Point List 22 Transformation Matrix (MIP) 23 Confidence measure 100 Data processing system (total system) 110 Input interface 120 Pre-registration module 130 Fine Registration Module (ICP) 135 Degrees of freedom module 140 Contact Optimization Module 150 output interface 160 Confidence Module S1--S3 Registration levels (coarse → fine → occlusal) K contact points P Penetration points VII. Brief description of the drawings Fig. 1 --- System block diagram: Data flow through modules 110--160 of the data processing system (100), from the separately recorded individual scans (10, 11) to the registered bite position (20) with transformation matrix (22), contact point list (21) and confidence measure (23). Fig. 2 --- Flowchart: Three-stage registration workflow S1 (pre-registration via dental arch geometry), S2 (fine registration using ICP) and S3 (contact optimization using evaluation function), including convergence check and confidence calculation. Fig. 3 --- Schematic representation of contact optimization: Left: Initial situation after fine-tuning with penetrations (P) and missing contacts. Right: Optimized MIP position with maximum contact points (K) at minimum penetration. Arrows show the optimization vector. VIII. Summary

[0017] Disclosed are a computer-implemented data processing system (claims 1-9), a registered dental bite position (claims 10-12) and a browser-based arrangement (claim 13) for automatically determining the maximum intercuspation position from separately recorded three-dimensional single-jaw geometries.

[0018] The system receives two separately acquired single-jaw scans, determines an initial spatial assignment based on geometric features of the dental arches, iteratively refines this assignment, and optimizes the occlusal contact situation by maximizing contact points while minimizing penetrations. The invention enables the determination of the bite relationship purely algorithmically from single-jaw scans—without a physical bite registration, without a buccal scan, and without manual assignment. IX. Short description (for DPMA form)

[0019] The invention relates to a computer-implemented system for automatically determining the maximum intercuspation position from separately acquired three-dimensional scans of individual jaws without a physical bite registration. Through multi-stage geometric registration and occlusal contact optimization, a reproducible bite position with a confidence level is determined, which serves as input for downstream analysis, simulation, or manufacturing systems. Reference symbol list 10 OK-Mesh (upper jaw single scan) 11 UK-Mesh (lower jaw single scan) 20 Registered Bite Positions (Issue) 21 Contact Point List 22 Transformation Matrix (MIP) 23 Confidence measure 100 Data processing system (total system) 110 Input interface 120 Pre-registration module 130 Fine Registration Module (ICP) 135 Degrees of freedom module 140 Contact Optimization Module 150 output interface 160 Confidence Module S1 Pre-registration (Preliminary registration) S2 Fine Registration (ICP) S3 Contact Optimization (Occlusal Optimization) K contact points P Penetration points

Claims

Computer-implemented data processing system for automatically determining a relative positional relationship between the upper and lower jaw, in particular a maximum intercuspal position (MIP), comprising: a) an input interface (110) configured to receive at least two separately acquired three-dimensional dental geometries of an upper and a lower jaw; b) a pre-registration module (120) configured to determine an initial spatial assignment of the individual jaw geometries to each other based on geometric features of the dental arches; c) a fine-registration module (130) configured for iterative refinement of the spatial assignment by minimizing a distance or correspondence measure between corresponding surface areas;> > d\) a contact optimization module (140) configured to determine > an optimal relative position by maximizing the number or > quality of occlusal contact points while minimizing > penetrations between the dental geometries; > > e\) an output interface (150) configured to provide > the determined relative position relationship as a transformation matrix, > contact point list and / or registered model pair for downstream > analysis, simulation or manufacturing systems.; Data processing system according to claim 1, wherein the three-dimensional dental geometries are available as triangulated surface meshes, as point cloud models or as volume models, in particular as separate intraoral scans of an upper and a lower jaw in STL, PLY, OBJ or a proprietary scanner format, which may have been acquired at different times or with different scanners. Data processing system according to claim 1 or 2, wherein the pre-registration module (120) determines the initial spatial assignment by means of at least one of the following methods: analysis of the dental arch geometry of both jaws to determine a plausible occlusal relation, recognition and assignment of anatomical landmarks such as cusp tips, fossae or marginal ridges, principal component analysis (PCA) to determine the principal axes of both dental arches, or bounding box alignment with subsequent transverse mirroring. Data processing system according to one of claims 1 to 3, wherein the fine registration module (130) employs an iterative closest point (ICP) method, a coherent point drift (CPD) registration algorithm, or any other iterative registration method that stepwise maximizes the agreement between corresponding surface areas of the maxillary and mandibular geometry. Data processing system according to one of claims 1 to 4, wherein the contact optimization module (140) calculates a multi-criteria evaluation function that aggregates at least two of the following criteria: number of occlusal contact points within a predetermined distance threshold, sum or maximum of the penetration depths between the geometries, spatial distribution of the contact points over the dental arch, and plausibility of the vertical bite height with respect to anatomical limits. Data processing system according to one of claims 1 to 5, wherein the system further comprises a degree-of-freedom module (135) that restricts the registration to clinically relevant degrees of freedom, in particular to vertical translation (bite height), horizontal translation (anterior-posterior and transverse) and rotation about the vertical axis, while tilting and rotations about the transverse and sagittal axes are restricted or fixed. Data processing system according to one of claims 1 to 6, wherein the system further comprises a confidence module (160) configured to calculate a registration quality or a confidence measure, in particular based on at least one of the following criteria: ratio of contact points to penetrations, convergence behavior of the iterative registration, comparison with statistical expected values ​​for occlusal contact patterns, or dispersion in multiple registrations of the same model pair with different starting values. Data processing system according to one of claims 1 to 7, wherein the system further comprises a normalization module or is coupled with a preceding normalization system that transforms the received single jaw geometries into a normalized coordinate system with respect to an automatically determined occlusal plane prior to registration and optionally extracts functional occlusal surfaces by means of direction-dependent filtering. Data processing system according to one of claims 1 to 8, wherein the system is configured to process multiple model pairs of the same patient, wherein a separate registration is performed for each pair and the system calculates an averaged optimal positional relationship and a dispersion as a confidence measure from the resulting registrations. A registered dental bite position, stored on a computer-readable medium, comprising: a machine-readable data set representing the spatial relationship between a three-dimensional maxillary geometry and a three-dimensional mandibular geometry in maximum intercuspation position, wherein the spatial relationship has been determined by means of automatic geometric registration from separately acquired single-jaw scans without taking a physical bite registration between the patient's jaws, and wherein the data set includes at least a transformation matrix and a list of occlusal contact points with associated distance values. Registered dental bite position according to claim 10, wherein the data set is further assigned a confidence measure that quantifies the registration quality, and wherein the data set is configured as input for at least one of the following downstream systems: a system for the inverse determination of temporomandibular joint parameters (CondylaMap), a system for movement simulation (condylar.pro), a system for automatic occlusal surface synthesis (CrownForge), a system for orthodontic analysis (OrthoGuide), or a system for the fabrication of dental restorations. Registered dental bite position according to claim 10 or 11, wherein the data set further comprises a contact map which contains a distance value to the antagonist for each discrete position on the occlusal surface, and wherein the contact map serves as input geometry for an evaluation module for determining articulation parameters. Arrangement for the registration-free determination of a dental bite position, comprising a data processing system according to one of claims 1 to 9, wherein the arrangement is executable as a browser-based application on a client computer and all processing steps are carried out client-side, so that the dental geometries do not leave the user's computer, or wherein the arrangement is installed as a local application on an end device.

Images