Method for generating learning model, diagnostic support system, and method for acquiring data

The image processing and data acquisition methods, combined with a learning model, enhance the diagnosis of nasal and paranasal sinus diseases by accurately differentiating conditions like odontogenic maxillary sinusitis, eosinophilic sinusitis, and others, facilitating early and effective treatment.

JP7725783B2Active Publication Date: 2025-08-20UNIVERSITY OF FUKUI
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
JP2024014603
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2024-02-02
Publication Date
2025-08-20
Estimated Expiration
2040-03-18

AI Technical Summary

Technical Problem

Existing technologies lack a diagnostic support system or method for accurately determining nasal and paranasal sinus diseases, particularly odontogenic maxillary sinusitis, eosinophilic sinusitis, sinus fungal infection, inverted papilloma, and maxillary cancer, using CT images.

Method used

An image processing method that selects and synthesizes CT images of the frontal, ethmoid, and sphenoid sinuses, combined with a data acquisition method that quantifies the relationship between these sinuses and their contents, and a learning model trained on CT image data and medical interview data to diagnose nasal and paranasal sinus diseases.

Benefits of technology

Enables accurate differentiation and early detection of nasal and paranasal sinus diseases, reducing the risk of delayed treatment and improving patient outcomes by providing a comprehensive diagnostic support system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007725783000002
    Figure 0007725783000002
  • Figure 0007725783000003
    Figure 0007725783000003
  • Figure 0007725783000004
    Figure 0007725783000004
Patent Text Reader

Abstract

To provide a system which assists diagnosis of a nasal and paranasal sinus disease.SOLUTION: A system comprises: an image processing device which divides a CT image of the maxillary sinus, the frontal sinus, the ethmoid sinus and the sphenoid sinus and a CT image including the bones around the maxillary sinus into five regions of the maxillary sinus, the frontal sinus, the front ethmoid sinus, the rear ethmoid sinus and the sphenoid sinus, specifies a region of objects generated therein and specifies the bones around the maxillary sinus; a data calculation device which numerizes a relation between the size of the five regions and the size of the regions of the objects generated therein, numerizes a difference between one bone and the other bone around the maxillary sinus and numerizes a concentration difference of the five regions; and a learning model device which determines a nasal and paranasal sinus disease by inputting the values obtained by the data calculation device to a learning model that has learned a region among the value obtained by numerizing the relation between the five regions and the objects generated therein, the value obtained by numerizing the difference between the one bone and the other bone around the maxillary sinus, the value obtained by numerizing the concentration difference of the five regions and the nasal and paranasal sinus disease of a person corresponding to the CT images.SELECTED DRAWING: Figure 19
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Knowledge-based segmentation of attenuation-related regions in the head

    JP2012531229A

  • Apparatus and method for treating sinus diseases

    JP2013514829A

  • Diagnosis support system using panoramic x-ray picture, and diagnosis support program using panoramic x-ray picture

    JP2015178019A

  • Automatic ent surgery preplanning using backtracking maze problem solution

    JP2017042616A

  • Medical image processor, medical image processing method, and medical image processing program

    JP2019208620A