patent.fyi
Applications  /  US 2026/0354967
United States Patent Application Publication

US 2026/0354967 A1

Published
Nov 12, 2026
Filed
May 8, 2025
Earliest date on the face
May 8, 2025
Classification
A61F 2/186
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Contact impression captures facial contours

What ifWhat if a contact impression captures the facial contours used to shape the prosthesis?
Claim language“a high-resolution 3D scan”

A skin-contact impression material records the contours around the defect and is backed by a rigid support shell. A gypsum cast made from the impression provides the surface on which a wax prosthesis pattern is shaped and its margins are trimmed.

1.2Manual annotation of facial detail

What ifWhat if a technician identifies facial surface features by inspecting the scan and aligned close-up photographs?
Claim language“using artificial intelligence”

The technician aligns the photographs to the surface mesh using visible facial landmarks. The technician traces wrinkles and scars on the mesh and marks visible pores as points, recording widths or depths where the scan resolves them.

1.3Measured geometry fills scan gaps

What ifWhat if measured geometry from the opposite side of the face supplies the missing surface?
Claim language“using trained anatomical inference models”

Mesh software reflects the intact region across a plane defined by midline facial landmarks. The operator aligns the reflected patch to the measured rim of the gap and blends the boundary so that both position and slope are continuous.

1.4Fitting during measured facial movement

What ifWhat if the patient moves their face while a trial prosthesis is fitted and observed?
Claim language“simulating motion”

The patient wears a silicone trial piece secured at the intended attachment locations and performs facial movements relevant to the fitting region. Close-up video records edge lift against a millimeter reference, and the fitter adjusts margin thickness or attachment placement before repeating the fitting.

1.6Silicone casting in a mold

What ifWhat if a two-part mold shapes the silicone prosthesis?
Claim language“directly printing the model”

A rigid split mold contains the contact surface and the textured outer surface, with vents at the ends of thin margins. A vacuum-degassed, two-part silicone mixture fills the closed mold and cures before the part is released and trimmed.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Silicone facial appliances for stage performance

A structured-light scanner records the performer’s face at rest and during the expressions required for a scene. A trained surface model labels creases and scar boundaries, and an artist-defined outer mesh sets the character geometry. A deformation solver uses the recorded expressions to position the appliance perimeter and adjust wall thickness where the face stretches. A dispensing printer deposits two-part silicone from the resulting model.

Component substitution

Poisson reconstruction for scan gap repair

Screened Poisson surface reconstruction replaces trained anatomical inference for small gaps in the facial scan. The reconstruction uses oriented points from the intact surface around each gap to generate a continuous patch. Software trims the patch to the gap boundary, joins it to the measured mesh, and retains the original surface samples elsewhere. The repaired mesh feeds the fit simulation and the silicone print model.

Parameter or scale variation

Variable layer heights across the prosthesis

A silicone dispensing printer deposits the prosthesis in 0.20 mm layers through the central body and 0.10 mm layers through the build heights that contain the tapered perimeter. A slicer selects the transition heights from the model and scales deposited volume per unit path length with layer height to maintain a constant nominal bead width. A removable support bath holds the tapered perimeter during deposition and cure.

Combination with existing systems

Prosthesis registration to bone anchored attachments

An optical scan captures the face and removable scan markers seated on existing bone-anchored attachment posts. Software matches the marker geometry to a component library and transfers the attachment positions and axes into the facial model. The print model includes pockets oriented to those axes, and a positioning fixture holds mechanical retention housings in the pockets while the silicone bond cures. The motion simulation constrains the attachment locations and computes strain in the surrounding silicone as the face moves.

Spectral skin mapping for silicone printing

A calibrated multispectral camera captures skin reflectance under controlled illumination and registers the image to the facial scan. Software estimates the color field from the surrounding skin and maps it onto the visible prosthesis surface. A bank of dispensing heads deposits compatible silicone formulations with calibrated pigment concentrations, and the slicer assigns a formulation to each surface path. An unpigmented silicone head deposits the interior using the same curing chemistry.

Automated or data-driven variant

Measured expressions drive thickness optimization

A synchronized camera array captures facial motion during smiling, blinking, and jaw opening. Digital image correlation estimates surface displacement, and registration maps those measurements onto the resting facial mesh. A finite element solver applies the measured displacements while an optimizer adjusts local silicone thickness to reduce perimeter lift under a minimum printable thickness constraint. The slicer converts the optimized geometry into deposition paths.

Specific version of a part

Separate surface maps guide deposition

The contact-surface height map, thickness map, and fine-relief map share an XY grid for a shallow prosthesis with no undercuts in the build direction. For each grid location, the controller reads the three values and deposits silicone only at layer heights between the contact surface and the outer surface calculated from thickness and relief.

Cross polarized structured light scanning

Two calibrated cameras record phase-shifted light patterns projected onto the face from several viewpoints. Crossed polarizers on the projector and camera lenses suppress surface glare, and the camera geometry converts pattern displacement into surface coordinates.

Neural labeling of surface features

A convolutional segmentation network receives aligned color and depth images from the scan. The network assigns pore, wrinkle, scar, or background labels to image pixels, and a calibrated projection transfers those labels onto the facial mesh.

Finite element simulation of fit

A finite element model represents the silicone with a hyperelastic material law fitted to tensile test data. Measured facial displacements drive the contact surface through selected expressions, and predicted edge separation and attachment forces guide adjustments to margin thickness and attachment placement.

Direct extrusion of mixed silicone

Two metering pumps feed a two-part, addition-cure silicone formulated for skin contact into a static mixer at the specified mix ratio. A gantry extrudes the mixed elastomer in successive layers onto a flat release sheet, and the printed prosthesis cures in place.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 11,704,901 B2
Published
Jul 18, 2023
Filed
Oct 17, 2022
Class
G06V 10/82
Document BUS 2024/0221150 A1
Published
Jul 4, 2024
Filed
Mar 20, 2024
Class
G06T 7/001
Claim 1
Document A
Document B
1.1(a) acquiring a high-resolution 3D scan of a patient's face;
1.2(b) processing the scan using artificial intelligence to identify facial surface features including pores, wrinkles, and scars;
A wrinkle detection service providing server for providing a wrinkle detection method based on an artificial intelligence, comprising: a data pre-processor for obtaining a skin image of a user from a skin measurement device and performing pre-processing based on feature points based on the skin image; a wrinkle detector for inputting the skin image pre-processed through the data pre-processing into an artificial neural network and generating a wrinkle probability map correspo
A · US 11,704,901 B2 · claim 1
Features of images of the cells with wrinkled electrode plates and the images of the cells with non-wrinkled electrode plates, such as wrinkle width and wrinkle texture of a wrinkled region, and wrinkling of separator, stripe pattern on separator, and stain on separator of a non-wrinkled region are extracted, and the convolutional neural network is used for repeated training to generate an AI deep learning electrode plate wrinkling detection model.
B · US 2024/0221150 A1 · ¶ 76
1.3(c) correcting deformities or scan gaps using trained anatomical inference models;
1.4(d) simulating motion of the facial region to optimize prosthetic fit and attachment;
1.5(e) generating a 3D model incorporating said detail; and
1.6(f) directly printing the model using biocompatible elastomeric silicone without use of any mold.

An empty cell means only that no passage was reproduced there.

Statements in these documents
There is increasing demand for a technique for diagnosing a user's skin condition based on a skin image or skin wrinkle image taken before using a skin improvement product and evaluating the efficacy of the skin improvement product based on a skin image or skin wrinkle image taken after using the skin improvement product.
A · US 11,704,901 B2 · ¶ 4
Therefore, the present disclosure has been made in view of the above problems, and it is an object of the present disclosure to provide a method of detecting wrinkles using a semantic segmentation model, a patch segmentation model, and an ensemble model combining the semantic segmentation model and the patch segmentation model.
A · US 11,704,901 B2 · ¶ 10
This detection method has the problems of lagged detection results and low accuracy of the detection results.
B · US 2024/0221150 A1 · ¶ 3
In addition, this also avoids false negatives and false positives caused by observation on the appearance of the cells and the problem of loss of manpower and material resources caused by disassembly of the normal cells due to disassembly of the cells for determining whether the cells have wrinkled electrode plates.
B · US 2024/0221150 A1 · ¶ 7
Cite aspatent.fyi, US 2026/0354967 A1, claim 1, first published 2026-11-19T06:07:06-05:00, https://patent.fyi/US20260354967/, SHA-256 d8fce350555c0ef74b3975df44d3bb187c9d2b07303388b469a687691e5c3fc8
First published
Permanent addresshttps://patent.fyi/US20260354967/
Content fingerprint (SHA-256)d8fce350555c0ef74b3975df44d3bb187c9d2b07303388b469a687691e5c3fc8
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi