20240008955. Automated Processing of Dental Scans Using Geometric Deep Learning simplified abstract (3M INNOVATIVE PROPERTIES COMPANY)

Automated Processing of Dental Scans Using Geometric Deep Learning

Organization Name

3M INNOVATIVE PROPERTIES COMPANY

Inventor(s)

Jonathan D. Gandrud of Woodbury MN (US)

Alexandra R. Cunliffe of St. Paul MN (US)

James D. Hansen of White Bear Lake MN (US)

Cameron M. Fabbri of St. Paul MN (US)

Wenbo Dong of Lakeville MN (US)

En-Tzu Yang of Cupertino CA (US)

Jianbing Huang of Shoreview MN (US)

Himanshu Nayar of Mountain View CA (US)

Guruprasad Somasundaram of Plymouth MN (US)

Jineng Ren of Fuzhou City (CN)

Joseph C. Dingeldein of Hudson WI (US)

Seyed Amir Hossein Hosseini of St. Paul MN (US)

Steven C. Demlow of Mendota Heights MN (US)

Benjamin D. Zimmer of Hudson WI (US)

Automated Processing of Dental Scans Using Geometric Deep Learning - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240008955 titled 'Automated Processing of Dental Scans Using Geometric Deep Learning

Simplified Explanation

The abstract of the patent application describes the application of machine learning, specifically geometric deep learning, to various dental processes. It mentions 5 specific solutions that utilize machine learning techniques, including generative adversarial networks for smile design, vertex and edge classification for gum versus teeth detection and teeth type segmentation, regression for coordinate systems, diagnostics, case complexity, and treatment duration prediction, and automatic encoders and clustering for grouping of doctors or technicians and preferences.

• Machine learning techniques, such as generative adversarial networks, are applied to smile design, appliance rendering, scan cleanup, restoration appliance design, crown and bridges design, and virtual debonding.
• Vertex and edge classification techniques are used to detect gums versus teeth, segment teeth types, and identify brackets and other orthodontic hardware.
• Regression techniques are applied to coordinate systems, diagnostics, case complexity, and prediction of treatment duration.
• Automatic encoders and clustering techniques are used to group doctors or technicians based on their preferences.

Potential applications of this technology:

• Improved smile design and appliance rendering for dental patients.
• Enhanced scan cleanup and restoration appliance design.
• More accurate and efficient crown and bridges design.
• Virtual debonding for orthodontic treatment planning.
• Automated detection and segmentation of gums, teeth types, and orthodontic hardware.
• Predictive models for treatment duration and case complexity.

Problems solved by this technology:

• Manual smile design and appliance rendering can be time-consuming and subjective. Machine learning can automate and improve this process.
• Accurate detection and segmentation of gums, teeth types, and orthodontic hardware can aid in treatment planning and diagnosis.
• Predictive models for treatment duration and case complexity can help in treatment planning and resource allocation.

Benefits of this technology:

• Improved efficiency and accuracy in dental processes, leading to better patient outcomes.
• Automation of repetitive tasks, allowing dental professionals to focus on more complex aspects of treatment.
• Enhanced treatment planning and diagnosis through automated detection and segmentation.
• Predictive models for treatment duration and case complexity can aid in resource allocation and scheduling.

Original Abstract Submitted

machine learning, or geometric deep learning, applied to various dental processes and 5 solutions. in particular, generative adversarial networks apply machine learning to smile design—finished smile, appliance rendering, scan cleanup, restoration appliance design, crown and bridges design, and virtual debonding. vertex and edge classification apply machine learning to gum versus teeth detection, teeth type segmentation, and brackets and other orthodontic hardware. regression applies machine learning to coordinate systems, diagnostics, case complexity, and 0 prediction of treatment duration. automatic encoders and clustering apply machine learning to grouping of doctors, or technicians, and preferences.