Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Single-step deep network-based automatic detection method for median multiple raw teeth in curved surface body layer graph

A deep network and automatic detection technology, applied in neural learning methods, biological neural network models, image analysis, etc., can solve problems such as errors in clinical diagnosis and data labeling, loss of global information, etc., achieve optimal prediction accuracy and reduce manual errors Effect

Pending Publication Date: 2022-05-27
NANJING UNIV OF POSTS & TELECOMM
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The size of the cropped image block is also manually set according to experience, not according to the image content, and errors often occur in the process of clinical diagnosis and data labeling
In addition, inputting image patches containing only local information into the training and testing stages will lose key global information

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Single-step deep network-based automatic detection method for median multiple raw teeth in curved surface body layer graph
  • Single-step deep network-based automatic detection method for median multiple raw teeth in curved surface body layer graph
  • Single-step deep network-based automatic detection method for median multiple raw teeth in curved surface body layer graph

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] like figure 1 As shown, an automatic detection method for median supernumerary teeth in surface tomography based on single-step deep network, including:

[0056] Acquire new surface layer images;

[0057] Input the new surface body layer image into the single-step deep network model obtained by training, and obtain the positions of multiple candidate bounding boxes and the target confidence that the target category belongs to the median supernumerary tooth;

[0058] The final bounding box was screened from multiple candidate bounding boxes using the non-maximum suppression method to obtain the final positioning of the median supernumerary teeth.

[0059] Further, in this embodiment, a single-step deep network model is obtained by training, including:

[0060] (1) Expand the training set, the training set includes surface tomography images, cone beam CT images, classification and annotation information and positioning annotation information;

[0061] (2) A single-ste...

Embodiment 2

[0094] For a surface layer image to be detected, in practical application, the following steps are specifically included:

[0095] (1) The training set is the surface tomography images, classification annotation information and positioning annotation information of 342 study cases, of which the size of the surface tomography images used for training is 2976×1536 and the pixel size is 0.07×0.07 mm 2 .

[0096] (2) Judging whether there is a median supernumerary tooth from the cone beam CT image corresponding to the curved tomographic image study case, as the classification and labeling information of the corresponding curved tomographic image. The curved tomographic images of the study cases were annotated, and the median supernumerary teeth in the training set of the curved tomographic images were used as the region of interest ROI, which was marked with a bounding box. remember R i for the surface layer image i ROI, extract all the images in the surface tomography R i...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a curved surface body layer image median multiple tooth automatic detection method based on a single-step deep network, and the method is characterized in that the method comprises the steps: collecting a new curved surface body layer image; inputting the new curved surface body layer image into a single-step deep network model obtained by training, and obtaining a plurality of candidate bounding box positions and a target confidence coefficient which contains a target category and belongs to a median multi-generation tooth; and screening out a final bounding box from the plurality of candidate bounding boxes by using a non-maximum suppression method to obtain final positioning of the median polygenic teeth. According to the method, only one-time scanning is carried out on the image of the curved surface body layer, whether the image contains the median polygenic teeth or not can be quickly and automatically identified, and the position of the median polygenic teeth is given, so that the influence of experience difference of doctors on precision is avoided, and the doctors are assisted in quickly and correctly diagnosing the median polygenic teeth.

Description

technical field [0001] The invention relates to an automatic detection method for supernumerary teeth in the middle of a curved body layer map based on a single-step deep network, and belongs to the technical field of computer vision prediction. Background technique [0002] Supernumerary teeth refer to the teeth or dentition-like tissues existing in the jawbone except for 20 of the deciduous dentition or 32 of the permanent dentition. It is a common disease of abnormal number of teeth. Supernumerary teeth often cause complications such as cystic lesions, delayed eruption of adjacent teeth, ectopic eruption of adjacent teeth, root resorption of adjacent teeth or torsion of adjacent teeth. Most supernumerary teeth are buried in the jawbone, so two-dimensional curved tomographic images are a common method for diagnosing and locating supernumerary teeth. Compared with cone beam computed tomography (Cone BeamComputed Tomography, CBCT), it has the advantages of small radiation d...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/00G06V10/774G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06T7/0012G06N3/04G06N3/08G06T2207/10081G06T2207/20081G06T2207/20084G06T2207/20104G06T2207/30036G06F18/214G06F18/24
Inventor 戴修斌蒋昕朱书进冒添逸刘天亮
Owner NANJING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com