CN113425440A - System and method for detecting caries and position thereof based on artificial intelligence - Google Patents
System and method for detecting caries and position thereof based on artificial intelligence Download PDFInfo
- Publication number
- CN113425440A CN113425440A CN202110708437.0A CN202110708437A CN113425440A CN 113425440 A CN113425440 A CN 113425440A CN 202110708437 A CN202110708437 A CN 202110708437A CN 113425440 A CN113425440 A CN 113425440A
- Authority
- CN
- China
- Prior art keywords
- caries
- tooth
- lesion
- module
- intelligent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61C—DENTISTRY; APPARATUS OR METHODS FOR ORAL OR DENTAL HYGIENE
- A61C19/00—Dental auxiliary appliances
- A61C19/04—Measuring instruments specially adapted for dentistry
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
- A61B5/0082—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence adapted for particular medical purposes
- A61B5/0088—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence adapted for particular medical purposes for oral or dental tissue
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30036—Dental; Teeth
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Medical Informatics (AREA)
- Veterinary Medicine (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Surgery (AREA)
- Artificial Intelligence (AREA)
- Heart & Thoracic Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pathology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Dentistry (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Epidemiology (AREA)
- Mathematical Physics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Dental Tools And Instruments Or Auxiliary Dental Instruments (AREA)
- Endoscopes (AREA)
Abstract
The invention relates to the technical field of caries detection, in particular to a system and a method for detecting caries and positions thereof based on artificial intelligence, wherein an optical filter module comprises: the device is used for filtering reflected light and scattered light of the light source module, so that excitation light of a new waveband can pass through the device conveniently; an image acquisition module: the tooth image acquisition device is used for acquiring tooth images in an exciting light state; the intelligent decayed tooth identification module: identifying the tooth image acquired by the image acquisition module to identify one or more positions of decayed teeth, lesion degrees and lesion boundaries; the invention filters scattered light and reflected light of an incident light source through the optical filter module, so that the interference of a tooth exciting light image is greatly reduced, the characteristics of lesions can be more obviously highlighted, meanwhile, the exciting light image marked with the features of the lesions is trained and learned on a large scale, an intelligent decayed tooth detection model is obtained, the traditional artificial detection is broken away, the automatic detection of the decayed tooth is realized, and the detection efficiency of the decayed tooth is improved.
Description
Technical Field
The invention relates to the technical field of caries detection, in particular to a system and a method for detecting caries and positions thereof based on artificial intelligence.
Background
Caries is a common, multifactorial, induced disease, primarily due to interactions of sugars and microorganisms on the tooth surface over time, involving lesions of dentin, cementum, and possibly further invading the pulp, causing pulpitis or periapical inflammation, or tooth loss. According to the 4 th national oral health epidemiological survey result, the following results are shown: the caries rate of permanent teeth of children aged 12 years is 34.5%, and the caries rate of deciduous teeth of children aged 5 years is 70%, and the caries rate is gradually increased, so that the caries is a common disease with high incidence. Therefore, detection of caries at an early stage and immediate intervention is a medical problem that is urgently needed to be solved.
The traditional method for diagnosing decayed teeth is mainly diagnosed by clinicians according to colors, shapes and qualities, lacks objective and uniform standards, and is difficult to provide a method for evaluating decayed teeth exactly. This subjective evaluation method according to doctors mainly has 2 problems: the subjective evaluations of different doctors are different; the subjective judgment standard of the same doctor is easy to have unstable judgment due to environmental reasons and the like. And diagnosis of early caries is more difficult. The cross-sectional photomicrography technique is the most practical and widely accepted standard for quantitatively researching early caries lesion gold at present. However, conventional imaging examinations have limited sensitivity and specificity for alveolar caries.
Hospital studies have shown that under UV light, teeth fluoresce spontaneously green, while carious lesions emit less intense fluorescence than healthy teeth. Therefore, research has been conducted to detect the mineralization of early caries using the fluorescence-guided technique as a non-invasive diagnostic tool. Nevertheless, the correlation coefficient of the immunofluorescence technique for the diagnosis of alveolar caries is only 0.53(Lee HS, J Biomed Opt,2018,23(9):1-7) with histological examination as gold standard, and the judgment for deeper alveolar caries is more limited. The scheme of judging human eyes and machines by simple image characteristics, but not the intelligent scheme of machine learning, has certain defects, including 1) that tooth tissues are green and not beneficial to image analysis; 2) the ability to detect demineralization in the tooth still requires more clinical trials to verify that the clinical application of red fluorescence is in the research phase. (Chenyan, journal of International oral medicine 2019, 46 (6): 699-702). Therefore, the accuracy of the detection result is also controversial, and only as an auxiliary diagnostic tool, the doctor needs to combine subjectivity and objectivity to make an accurate judgment (Wangkexin, 2020,28 (11): 744-748).
In the current patent, there is a group developing a fluoroscope (see patent specifications CN205683069U, CN106821529A) which is convenient for ordinary people to use at home, and the device excites dental plaque on the tooth surface by a deep blue LED to generate red fluorescence, and automatically observes whether the tooth has dental plaque and the degree of the dental plaque. At present, a dental plaque fluorescence detector is developed by taking a laser as a light source and combining a camera technology for hospitals. However, this type of device is not useful for caries detection. The Netherlands Philips corporation developed an early caries detection method and apparatus (see patent specification CN110769746A) that uses the transmission of light through the tooth and the reflection of light from the tooth detected by an imaging device to detect caries. This is a measurement of caries using transmission and emission of light rather than detection of excitation light from the tooth as used herein, and in one aspect, does not use artificial intelligence to determine caries. X-ray imaging is another caries detection technique, but it is not suitable for portable home use. Another early caries detection technique was Fiber Optic Transmission Illumination (FOTI), which examines teeth by visible light illumination, but this technique suffers from scattered light from the teeth, resulting in poor contrast. In addition, there is a caries detection using infrared light, which is a detection using infrared light scattering properties of a carious portion. The existing caries detection method is not only expensive or based on the scattering/transmission/reflection phenomenon of the light source (instead of the exciting light mentioned in the patent), but also does not adopt artificial intelligence means, only can be used as auxiliary means, and further judgment is needed manually, therefore, the technical personnel in the field provide a system and a method for detecting caries and the position thereof based on artificial intelligence, so as to solve the problems in the background technology.
Disclosure of Invention
The present invention is directed to a system and method for detecting caries and its location based on artificial intelligence to solve the problems set forth in the background art.
In order to achieve the purpose, the invention provides the following technical scheme: a system for detecting caries and its location based on artificial intelligence, the system comprising:
a light source module: the tooth can be irradiated by light with the wavelength in the first wavelength range, the tooth can be excited to generate excitation light with a new waveband after absorbing light energy, and the wavelength of the excitation light is in the second wavelength range; the first wavelength range and the second wavelength range are different.
The optical filter module: the device is used for filtering reflected light and scattered light of the light source module, so that excitation light of a new waveband can pass through the device conveniently;
an image acquisition module: the tooth image acquisition device is used for acquiring tooth images in an exciting light state;
the intelligent decayed tooth identification module: identifying the tooth image acquired by the image acquisition module to identify one or more positions of decayed teeth, lesion degrees and lesion boundaries;
an image display module: adopting an intelligent terminal;
a transmission module: wireless communication technologies such as WIFI or Bluetooth or USB wired transmission are adopted, and the image display module is connected in a matched mode to realize data transmission and communication;
a power supply module: as a supply of electrical energy;
the intelligent caries labeling module: the system is used for marking the position, the lesion degree and the lesion boundary of the decayed tooth on a tooth image, outputting the position, the lesion degree and the lesion boundary of the decayed tooth to an image display module through a transmission module, and finally realizing intelligent automatic detection on the decayed tooth and the characteristics of the positions, the quantity, the lesion degree, the lesion boundary and the like of the decayed tooth;
in addition, the use method of the system for detecting caries and positions thereof based on artificial intelligence is as follows:
step 1: the light source module is aligned to the oral cavity of a patient, emits light with a first wavelength range to irradiate the teeth of the patient, the teeth absorb light energy, and excite new-waveband excitation light with a second wavelength range, the first wavelength range and the second wavelength range are different, after healthy teeth and lesions in the teeth absorb the light energy, different and composite characteristics including intensity, color, environmental information and the like can be optically shown, the filter module filters scattered and reflected light with the first wavelength range, so that the new-waveband excitation light with the second wavelength range can enter the image acquisition module, and further, the characteristics of caries lesion areas for acquiring pictures can be more obvious;
step 2: the image acquisition module transmits the acquired image to the intelligent decayed tooth identification module, the intelligent decayed tooth identification module is a detection model obtained by object based on a deep neural network and training of a positioning identification algorithm, the input data set carries out identification steps through the detection model, the identification steps include but are not limited to convolution, pooling and full connection, and the identification output comprises the position of the decayed tooth and the lesion degree;
and step 3: the intelligent caries labeling module can draw a boundary box around the caries, and label information such as lesion degree, lesion quantity, lesion tooth number and the like in a text form; for example: the lesion level is scored into 5 levels, with level a being the most severe and level E the least severe, and the 32 teeth of the person numbered sequentially from a to E, where teeth with caries are identified and the numbers are displayed;
and 4, step 4: the transmission module transmits information such as the degree of the marked lesions, the number of the lesions, the tooth numbers of the lesions and the like to the image display module and displays the information.
Preferably, in one embodiment, the first wavelength range is 390-430 nm;
preferably, in another embodiment, the first wavelength range is 620-1200 nm;
preferably, in another embodiment, the first wavelength range is formed by combining a plurality of wavelength ranges, such as 390-430nm and 620-1200 nm.
Preferably, in one embodiment, the second wavelength range is 450-600 nm;
preferably, in another embodiment, the second wavelength range is 620-750 nm;
preferably, in another embodiment, the second wavelength range is a combination of multiple long ranges, such as 450-600nm and 620-750 nm.
In one embodiment, the dental excitation light is a fluorescent light.
In one embodiment, the filter module filters out a part of the wavelength range of the light, and allows another part of the wavelength range of the light to pass through.
As a further aspect of the invention: the light source module is arranged on the periphery of the optical filter module, adopts a light emitting diode or a laser, and is convenient to be incident to the tooth and the image acquisition module as illumination.
As a further aspect of the invention: the optical filter module is arranged in front of the image acquisition module and used for shielding reflected light and scattered light with the wavelength within a first wavelength range, and the image acquisition module adopts a high-definition camera which is a CCD or CMOS chip.
As a further aspect of the invention: the intelligent decayed tooth identification module comprises a calculation unit and a storage unit, wherein the storage unit is used for storing tooth images, detection models, user data and the like, the calculation unit is used for intelligently identifying decayed teeth, and the identification method comprises the following steps:
s1: performing computer numerical operation on the input excitation light image to obtain multi-scale characteristics of the excitation light image, wherein the multi-scale characteristics comprise numerical values of red components, green components and blue components of each sub-pixel, distribution and average values of the numerical values, the distribution and the average values of the red components, the green components and the blue components of each sub-pixel, and the like;
s2: taking the multi-scale features of the excitation light image as independent variables, outputting the independent variables to a caries intelligent detection model for calculation, and selectively and automatically outputting caries related dependent variables, including lesion degree, lesion position and quantity of each tooth, area of each lesion region and boundary of the lesion region;
as a further aspect of the invention: the multi-scale characteristic related to the exciting light image and the dependent variable related to the caries present an n-element m-degree polynomial function relationship, the numerical value of the dependent variable related to the caries can be obtained by finite multiplication and addition operation of a constant and an independent variable, the polynomial function relationship and the constant are obtained by continuously optimizing the parameters of a classifier on the basis of a group of training samples marked with the caries characteristic, and the polynomial function relationship is as follows:
wherein x isnRepresenting a feature in the excitation light image, knCoefficients representing the pair of terms, dependent variable f (x) associated with caries1，x2,…,xn) When a certain threshold is reached, the area can be predicted as a caries lesion area.
As a further aspect of the invention: the caries intelligent detection model is obtained by a machine learning training according to the functional relation between a dependent variable related to an exciting light image and a caries related dependent variable, and the specific learning method is as follows:
firstly, artificially marking a decayed tooth lesion part, forming a large-scale decayed tooth characteristic training sample, preferentially processing a missing value and a repeated value of the decayed tooth characteristic training sample, and converting a data type;
then extracting and selecting the characteristics of the decayed tooth, calculating the correlation of each characteristic of the decayed tooth, determining a correlation coefficient, and selecting some characteristics as the input of an intelligent detection model of the decayed tooth according to the magnitude of the correlation coefficient;
selecting features and labels of caries training data and caries test data, dividing data greater than or equal to 50% into training data sets, simultaneously importing machine learning algorithms, and creating and training caries intelligent detection models, wherein the machine learning algorithms include but are not limited to random forests, support vector machines, naive Bayes, neural networks and the like.
As a further aspect of the invention: the boundary frame adopts any one of rectangular, circular or polygonal contours;
rectangle: the coordinates of the upper left corner and the coordinates of the lower right corner are determined, so that the boundary box can circle the continuous lesion areas;
circular: determining the coordinates of circle centers, wherein the circle takes the center of the lesion as the center, and if the identified lesion area is not continuous, performing frame selection by a plurality of circle centers and performing annotation in the surrounding area;
polygonal profile: and (3) performing contour drawing by a computer along the boundary of the lesion and the healthy tooth body, starting from an identified boundary point, taking the multi-scale features of the dental caries lesion as a criterion of the boundary, continuously searching, and outlining the lesion area in a frame drawing mode.
As a further aspect of the invention: the intelligent detection model for dental caries also comprises an intelligent identification model for dental cryptorrhoea or dental plaque, wherein the intelligent identification model for dental fissure or dental plaque is a detection model obtained by training an object identification and positioning identification algorithm based on a deep neural network, the input data set is subjected to the steps of convolution, pooling, full connection and the like through the detection model, and finally lesion information containing dental plaque and dental cryptorrhoea is output.
Compared with the prior art, the invention has the beneficial effects that: the invention filters scattered light and reflected light of an incident light source through the optical filter module, so that the interference of a tooth exciting light image is greatly reduced, the characteristics of lesions can be more obviously highlighted, meanwhile, the exciting light image marked with the features of the lesions is trained and learned on a large scale, an intelligent decayed tooth detection model is obtained, the traditional artificial detection is broken away, the automatic detection of the decayed tooth is realized, and the detection efficiency, the accuracy, the sensitivity and the specificity of the decayed tooth are improved.
Drawings
FIG. 1 is a schematic flow chart of a system for detecting caries and its location based on artificial intelligence;
FIG. 2 is a schematic diagram of the image acquisition module in a system and method for detecting caries and location based on artificial intelligence;
FIG. 3 is a schematic diagram of the operation of a dual light source image acquisition module in a system and method for detecting caries and its location based on artificial intelligence.
Detailed Description
Referring to fig. 1 to 3, in an embodiment of the present invention, a system for detecting caries and its location based on artificial intelligence includes:
a light source module: the light with the wavelength of 620-1200nm can be emitted to irradiate the tooth, and after the tooth absorbs the light energy, the tooth is excited to generate excitation light with a new wave band, wherein the wavelength of the excitation light is 450-600 nm;
the optical filter module: the device is used for filtering reflected light and scattered light of the light source module, so that excitation light of a new waveband can pass through the device conveniently;
an image acquisition module: the tooth image acquisition device is used for acquiring tooth images in an exciting light state;
the intelligent decayed tooth identification module: identifying the tooth image acquired by the image acquisition module to identify one or more positions of decayed teeth, lesion degrees and lesion boundaries;
an image display module: adopting an intelligent terminal;
a transmission module: wireless communication technologies such as WIFI or Bluetooth or USB wired transmission are adopted, and the image display module is connected in a matched mode to realize data transmission and communication;
a power supply module: as a supply of electrical energy;
the intelligent caries labeling module: the system is used for marking the position, the lesion degree and the lesion boundary of the decayed tooth on a tooth image, outputting the position, the lesion degree and the lesion boundary of the decayed tooth to an image display module through a transmission module, and finally realizing intelligent automatic detection on the decayed tooth and the characteristics of the positions, the quantity, the lesion degree, the lesion boundary and the like of the decayed tooth;
in addition, the use method of the system for detecting caries and positions thereof based on artificial intelligence is as follows:
step 1: the light source module is aligned to the oral cavity of a patient, emits light with the wavelength of 620-1200nm, irradiates the teeth of the patient, the teeth absorb light energy, and excite new-waveband excitation light with the wavelength of 450-1200 nm, after the healthy teeth and lesion parts in the teeth absorb the light energy, different and composite characteristics including intensity, color, environmental information and the like can be presented optically, the filter module filters the scattered and reflected light with the wavelength of 620-1200nm, so that only the new-waveband excitation light with the wavelength of 450-600nm can enter the image acquisition module, and further the characteristics of the carious lesion area of the acquired image can be more obvious;
step 2: the image acquisition module transmits the acquired image to the intelligent decayed tooth identification module, the intelligent decayed tooth identification module is a detection model obtained by object based on a deep neural network and training of a positioning identification algorithm, the input data set carries out identification steps through the detection model, the identification steps include but are not limited to convolution, pooling and full connection, and the identification output comprises the position of the decayed tooth and the lesion degree;
and step 3: the intelligent caries labeling module can draw a boundary box around the caries, and label information such as lesion degree, lesion quantity, lesion tooth number and the like in a text form; for example: the lesion level is scored into 5 levels, with level a being the most severe and level E the least severe, and the 32 teeth of the person numbered sequentially from a to E, where teeth with caries are identified and the numbers are displayed;
and 4, step 4: the transmission module transmits information such as the degree of the marked lesions, the number of the lesions, the tooth numbers of the lesions and the like to the image display module and displays the information.
Furthermore, the light source module is arranged on the periphery of the optical filter module, and the light source module adopts a light emitting diode or a laser, so that the light can be conveniently incident to the teeth and the image acquisition module as illumination.
Furthermore, the optical filter module is arranged in front of the image acquisition module and used for shielding reflected light and scattered light with the wavelength of 390-plus-430 nm, the image acquisition module adopts a high-definition camera, and the high-definition camera adopts a CCD or CMOS chip.
Further, the intelligent decayed tooth identification module comprises a computing unit and a storage unit, wherein the storage unit is used for tooth images, detection models, user data and the like, the computing unit is used for intelligent decayed tooth identification, and the identification method comprises the following steps:
s1: performing computer numerical operation on the input excitation light image to obtain multi-scale characteristics of the excitation light image, wherein the multi-scale characteristics comprise numerical values of red components, green components and blue components of each sub-pixel, distribution and average values of the numerical values, the distribution and the average values of the red components, the green components and the blue components of each sub-pixel, and the like;
s2: taking the multi-scale features of the excitation light image as independent variables, outputting the independent variables to a caries intelligent detection model for calculation, and selectively and automatically outputting caries related dependent variables, including lesion degree, lesion position and quantity of each tooth, area of each lesion region and boundary of the lesion region;
furthermore, the multi-scale features related to the excitation light image and the dependent variable related to caries are in an n-element m-degree polynomial function relationship, the numerical value of the dependent variable related to caries can be obtained by finite multiplication and addition operation of a constant and an independent variable, the polynomial function relationship and the constant are obtained by continuously optimizing the parameters of the classifier on the basis of a group of training samples marked with the caries features, and the polynomial function relationship is as follows:
wherein x isnRepresenting a feature in the excitation light image, knCoefficients representing the pair of terms, dependent variable f (x) associated with caries1，x2,…,xn) When a certain threshold is reached, the area can be predicted as a caries lesion area.
Furthermore, the caries intelligent detection model is obtained by machine learning training of a functional relation between a dependent variable related to an excitation light image and a caries related dependent variable, and the specific learning method is as follows:
firstly, artificially marking a decayed tooth lesion part, forming a large-scale decayed tooth characteristic training sample, preferentially processing a missing value and a repeated value of the decayed tooth characteristic training sample, and converting a data type;
then extracting and selecting the characteristics of the decayed tooth, calculating the correlation of each characteristic of the decayed tooth, determining a correlation coefficient, and selecting some characteristics as the input of an intelligent detection model of the decayed tooth according to the magnitude of the correlation coefficient;
selecting features and labels of caries training data and caries test data, dividing data greater than or equal to 50% into training data sets, simultaneously importing machine learning algorithms, and creating and training caries intelligent detection models, wherein the machine learning algorithms include but are not limited to random forests, support vector machines, naive Bayes, neural networks and the like.
Further, the bounding box adopts any one of rectangular, circular or polygonal outlines;
rectangle: the coordinates of the upper left corner and the coordinates of the lower right corner are determined, so that the boundary box can circle the continuous lesion areas;
circular: determining the coordinates of circle centers, wherein the circle takes the center of the lesion as the center, and if the identified lesion area is not continuous, performing frame selection by a plurality of circle centers and performing annotation in the surrounding area;
polygonal profile: and (3) performing contour drawing by a computer along the boundary of the lesion and the healthy tooth body, starting from an identified boundary point, taking the multi-scale features of the dental caries lesion as a criterion of the boundary, continuously searching, and outlining the lesion area in a frame drawing mode.
Furthermore, the intelligent detection model for dental caries also comprises an intelligent identification model for dental cryptorrhoea or dental plaque, wherein the intelligent identification model for dental fissure or dental plaque is a detection model obtained by training an object identification and positioning identification algorithm based on a deep neural network, the input data set is subjected to the steps of convolution, pooling, full connection and the like through the detection model, and finally lesion information containing dental plaque and dental cryptorrhoea is output.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention are equivalent to or changed within the technical scope of the present invention.
Claims (8)
1. A system for detecting caries and its location based on artificial intelligence, the system comprising:
a light source module: the tooth can be irradiated by light with the wavelength in the first wavelength range, the tooth can be excited to generate excitation light with a new waveband after absorbing light energy, and the wavelength of the excitation light is in the second wavelength range; the first wavelength range and the second wavelength range are different.
The optical filter module: the device is used for filtering reflected light and scattered light of the light source module, so that excitation light of a new waveband can pass through the device conveniently;
an image acquisition module: the tooth image acquisition device is used for acquiring tooth images in an exciting light state;
the intelligent decayed tooth identification module: identifying the tooth image acquired by the image acquisition module to identify one or more positions of decayed teeth, lesion degrees and lesion boundaries;
an image display module: adopting an intelligent terminal;
a transmission module: the image display module is connected with the wireless communication technology such as WIFI or Bluetooth or the USB wired communication technology in a matching way, so that data transmission and communication are realized;
a power supply module: as a supply of electrical energy;
the intelligent caries labeling module: the system is used for marking the position, the lesion degree and the lesion boundary of the decayed tooth on a tooth image, outputting the position, the lesion degree and the lesion boundary of the decayed tooth to an image display module through a transmission module, and finally realizing intelligent automatic detection on the decayed tooth and the characteristics of the positions, the quantity, the lesion degree, the lesion boundary and the like of the decayed tooth;
in addition, the use method of the system for detecting caries and positions thereof based on artificial intelligence is as follows:
step 1: the light source module is aligned to the oral cavity of a patient, emits light with a first wavelength range to irradiate the teeth of the patient, the teeth absorb light energy, and excite new-waveband excitation light with a second wavelength range, the first wavelength range and the second wavelength range are different, after healthy teeth and lesions in the teeth absorb the light energy, different and composite characteristics including intensity, color, environmental information and the like can be optically shown, the filter module filters scattered and reflected light with the first wavelength range, so that the new-waveband excitation light with the second wavelength range can enter the image acquisition module, and further, the characteristics of caries lesion areas for acquiring pictures can be more obvious;
step 2: the image acquisition module transmits the acquired image to the intelligent decayed tooth identification module, the intelligent decayed tooth identification module is a detection model obtained by object based on a deep neural network and training of a positioning identification algorithm, the input data set carries out identification steps through the detection model, the identification steps include but are not limited to convolution, pooling and full connection, and the identification output comprises the position of the decayed tooth and the lesion degree;
and step 3: the intelligent caries labeling module can draw a boundary box around the caries, and label information such as lesion degree, lesion quantity, lesion tooth number and the like in a text form; for example: the lesion level is scored into 5 levels, with level a being the most severe and level E the least severe, and the 32 teeth of the person numbered sequentially from a to E, where teeth with caries are identified and the numbers are displayed;
and 4, step 4: the transmission module transmits information such as the degree of the marked lesions, the number of the lesions, the tooth numbers of the lesions and the like to the image display module and displays the information.
2. The system and method for artificial intelligence based detection of caries and their location as claimed in claim 1 wherein the light source module is installed at the periphery of the filter module and light emitting diode or laser is used as the light source module for incidence to teeth and image acquisition module.
3. The system and method for artificial intelligence based detection of caries and their location as claimed in claim 1, wherein the filter module is placed in front of the image capture module for shielding the reflected and scattered light with the first wavelength range, the image capture module is a high definition camera, and the high definition camera is a CCD or CMOS chip.
4. The system and method for detecting caries and its position based on artificial intelligence as claimed in claim 1, characterized by that, the caries intelligent identification module includes a calculation unit and a storage unit, the storage unit is used for storing the caries tooth image, detection model, user data, etc., the calculation unit is used for caries intelligent identification, the identification method is as follows:
s1: performing computer numerical operation on the input excitation light image to obtain multi-scale characteristics of the excitation light image, wherein the multi-scale characteristics comprise numerical values of red components, green components and blue components of each sub-pixel, distribution and average values of the numerical values, the distribution and the average values of the red components, the green components and the blue components of each sub-pixel, and the like;
s2: the multi-scale features of the excitation light image are used as independent variables, calculation is carried out in an intelligent caries detection model, and caries related dependent variables including lesion degree, lesion position and quantity, area of each lesion region and boundaries of the lesion regions of each tooth are selectively and automatically output.
5. The system and method for detecting caries and their location based on artificial intelligence as claimed in claim 4, wherein the multi-scale features related to the excitation light image and the caries related dependent variable are in a polynomial function of degree n m, the values of caries related dependent variable can be obtained by finite multiplication and addition of constants and independent variables, the polynomial function and constants are obtained by continuously optimizing the parameters of the classifier based on a set of training samples labeled with caries features, and the polynomial function is as follows:
wherein x isnRepresenting a feature in the excitation light image, knCoefficients representing the pair of terms, dependent variable f (x) associated with caries1，x2,…,xn) When a certain threshold is reached, the area can be predicted as a caries lesion area.
6. The system and method for detecting caries and their location based on artificial intelligence as claimed in claim 4, wherein the caries intelligent detection model is obtained by machine learning training from the function relationship between the dependent variable related to the excitation light image and the dependent variable related to caries, and the specific learning method is as follows:
firstly, artificially marking a decayed tooth lesion part, forming a large-scale decayed tooth characteristic training sample, preferentially processing a missing value and a repeated value of the decayed tooth characteristic training sample, and converting a data type;
then extracting and selecting the characteristics of the decayed tooth, calculating the correlation of each characteristic of the decayed tooth, determining a correlation coefficient, and selecting some characteristics as the input of an intelligent detection model of the decayed tooth according to the magnitude of the correlation coefficient;
selecting features and labels of caries training data and caries test data, dividing data greater than or equal to 50% into training data sets, simultaneously importing machine learning algorithms, and creating and training caries intelligent detection models, wherein the machine learning algorithms include but are not limited to random forests, support vector machines, naive Bayes, neural networks and the like.
7. The system and method for artificial intelligence based detection of caries and their location as claimed in claim 1, wherein: the boundary frame adopts any one of rectangular, circular or polygonal contours;
rectangle: the coordinates of the upper left corner and the coordinates of the lower right corner are determined, so that the boundary box can circle the continuous lesion areas;
circular: determining the coordinates of circle centers, wherein the circle takes the center of the lesion as the center, and if the identified lesion area is not continuous, performing frame selection by a plurality of circle centers and performing annotation in the surrounding area;
polygonal profile: and (3) performing contour drawing by a computer along the boundary of the lesion and the healthy tooth body, starting from an identified boundary point, taking the multi-scale features of the dental caries lesion as a criterion of the boundary, continuously searching, and outlining the lesion area in a frame drawing mode.
8. The system and method for artificial intelligence based detection of caries and their location according to claim 6, wherein the intelligent caries detection model further comprises an intelligent identifying model for tooth hidden fissures or dental plaques, the intelligent identifying model for tooth hidden fissures or dental plaques is a detection model trained by object identification and positioning identification algorithm based on deep neural network, the input data set is convolved, pooled and fully connected by the detection model, and finally lesion information containing dental plaques and tooth hidden fissures is output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110708437.0A CN113425440A (en) | 2021-06-24 | 2021-06-24 | System and method for detecting caries and position thereof based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110708437.0A CN113425440A (en) | 2021-06-24 | 2021-06-24 | System and method for detecting caries and position thereof based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113425440A true CN113425440A (en) | 2021-09-24 |
Family
ID=77754274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110708437.0A Pending CN113425440A (en) | 2021-06-24 | 2021-06-24 | System and method for detecting caries and position thereof based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113425440A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115153900A (en) * | 2022-07-27 | 2022-10-11 | 宋海洋 | Dental caries removal method and dental caries removal system based on dental surgery robot |
| WO2023198101A1 (en) * | 2022-04-12 | 2023-10-19 | 先临三维科技股份有限公司 | Artificial intelligence-based oral cavity examination method and apparatus, electronic device, and medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101588750A (en) * | 2007-01-17 | 2009-11-25 | 卡尔斯特里姆保健公司 | System for early detection of dental caries |
| CN103442645A (en) * | 2011-03-21 | 2013-12-11 | 卡尔斯特里姆保健公司 | A method for tooth surface classification |
| CN107863149A (en) * | 2017-11-22 | 2018-03-30 | 中山大学 | A kind of intelligent dentist's system |
| CN108143396A (en) * | 2018-01-19 | 2018-06-12 | 苏州江奥光电科技有限公司 | A kind of portable early stage caries diagnostic device and diagnostic method based on fluorescence imaging |
| CN109893101A (en) * | 2019-04-26 | 2019-06-18 | 中国科学院长春光学精密机械与物理研究所 | A kind of caries diagnosis imaging device and hand-held caries diagnosis imager |
| CN111083922A (en) * | 2018-08-21 | 2020-04-28 | 数码牙科集线 | Dental image analysis method for correction diagnosis and apparatus using the same |
| CN111784639A (en) * | 2020-06-05 | 2020-10-16 | 浙江大学 | Oral panoramic film dental caries depth identification method based on deep learning |
| CN111973152A (en) * | 2020-06-17 | 2020-11-24 | 谈斯聪 | Five sense organs and surgical medical data acquisition analysis diagnosis robot and platform |
| CN112037913A (en) * | 2020-09-15 | 2020-12-04 | 南昌大学 | Intelligent periodontitis detection method and system based on convolutional neural network |
| CN112150422A (en) * | 2020-09-15 | 2020-12-29 | 苏州知会智能科技有限公司 | Modeling method of oral health self-detection model based on multitask learning |
-
2021
- 2021-06-24 CN CN202110708437.0A patent/CN113425440A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101588750A (en) * | 2007-01-17 | 2009-11-25 | 卡尔斯特里姆保健公司 | System for early detection of dental caries |
| CN103442645A (en) * | 2011-03-21 | 2013-12-11 | 卡尔斯特里姆保健公司 | A method for tooth surface classification |
| CN107863149A (en) * | 2017-11-22 | 2018-03-30 | 中山大学 | A kind of intelligent dentist's system |
| CN108143396A (en) * | 2018-01-19 | 2018-06-12 | 苏州江奥光电科技有限公司 | A kind of portable early stage caries diagnostic device and diagnostic method based on fluorescence imaging |
| CN111083922A (en) * | 2018-08-21 | 2020-04-28 | 数码牙科集线 | Dental image analysis method for correction diagnosis and apparatus using the same |
| CN109893101A (en) * | 2019-04-26 | 2019-06-18 | 中国科学院长春光学精密机械与物理研究所 | A kind of caries diagnosis imaging device and hand-held caries diagnosis imager |
| CN111784639A (en) * | 2020-06-05 | 2020-10-16 | 浙江大学 | Oral panoramic film dental caries depth identification method based on deep learning |
| CN111973152A (en) * | 2020-06-17 | 2020-11-24 | 谈斯聪 | Five sense organs and surgical medical data acquisition analysis diagnosis robot and platform |
| CN112037913A (en) * | 2020-09-15 | 2020-12-04 | 南昌大学 | Intelligent periodontitis detection method and system based on convolutional neural network |
| CN112150422A (en) * | 2020-09-15 | 2020-12-29 | 苏州知会智能科技有限公司 | Modeling method of oral health self-detection model based on multitask learning |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023198101A1 (en) * | 2022-04-12 | 2023-10-19 | 先临三维科技股份有限公司 | Artificial intelligence-based oral cavity examination method and apparatus, electronic device, and medium |
| CN115153900A (en) * | 2022-07-27 | 2022-10-11 | 宋海洋 | Dental caries removal method and dental caries removal system based on dental surgery robot |
| CN115153900B (en) * | 2022-07-27 | 2024-01-12 | 赫丽佰(合肥)智能科技有限公司 | Caries removing method and caries removing system based on dental surgical robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10251558B2 (en) | Method and apparatus for measuring and displaying dental plaque, gingiva, and alveolar bone | |
| US9870613B2 (en) | Detection of tooth condition using reflectance images with red and green fluorescence | |
| CN103549938B (en) | Method for a real-time visualization of caries condition | |
| Rodrigues et al. | Light-emitting diode and laser fluorescence–based devices in detecting occlusal caries | |
| US20220189611A1 (en) | Noninvasive multimodal oral assessment and disease diagnoses apparatus and method | |
| JP2017537744A (en) | Intraoral 3D fluorescence imaging | |
| US20100279248A1 (en) | Device and method for predicting the likelihood of caries development | |
| US20040202356A1 (en) | Optical detection of dental caries | |
| MXPA04011069A (en) | System and method for detecting dental caries. | |
| Bussaneli et al. | Assessment of a new infrared laser transillumination technology (808 nm) for the detection of occlusal caries—An in vitro study | |
| JP2007505645A (en) | Automated endoscope device, diagnostic method and usage | |
| Kim et al. | A new screening method to detect proximal dental caries using fluorescence imaging | |
| CN113425440A (en) | System and method for detecting caries and position thereof based on artificial intelligence | |
| WO2011140663A1 (en) | Method of processing and displaying oral health diagnostic data | |
| Bansode et al. | Diagnosing dental caries: An insight | |
| Keem et al. | Wavelet representations for monitoring changes in teeth imaged with digital imaging fiber-optic transillumination | |
| Salsone et al. | Histological validation of near-infrared reflectance multispectral imaging technique for caries detection and quantification | |
| Presoto et al. | Clinical effectiveness of fluorescence, digital images and ICDAS for detecting occlusal caries | |
| Gawad et al. | Classification of dental diseases using hyperspectral imaging and laser induced fluorescence | |
| CN111798445A (en) | Tooth image caries identification method and system based on convolutional neural network | |
| Mital et al. | Recent advances in detection and diagnosis of dental caries | |
| Mahmoud et al. | Instant identification of dental white spot using K-means algorithm via laser-induced fluorescence and associated hyperspectral imaging | |
| Simon et al. | Assessment of cavitation in artificial approximal dental lesions with near-IR imaging | |
| Kim et al. | Quantitative light-induced fluorescence technology for quantitative evaluation of tooth wear | |
| Wang et al. | Machine learning-based automatic identification and diagnosis of dental caries and calculus using hyperspectral fluorescence imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |