KR20160112494A - Effective oral lesion onset detection method and system in Panorama - Google Patents

Abstract

The present invention is provided for exploring oral lesion to predict and sense the same. The present invention relates to a method for predicting and sensing whether oral lesion occurs, which is configured to set an exploration area by using a feature vector of an area case, and the like with respect to a panorama image of an oral cavity, to emphasize the set exploration area to explore a lesion position in the panorama image of a patient by using information on a lesion case, and to accurately explore lesion while quickly performing lesion exploration. The method for predicting and sensing whether oral lesion occurs using an oral lesion occurrence predicting system comprises: an exploration area setting step of setting an exploration area for exploring lesion in a candidate image of a panorama image in contrast with area case information loaded from a data storage unit; a lesion exploration step of loading a lesion case image for exploring lesion from the data storage unit, and detecting a lesion candidate area which is contrasted with the lesion case image from the exploration area set in the exploration area setting step; and a lesion display step of displaying image information on the lesion candidate area explored in the lesion exploration step, corresponding case image information, information data of the corresponding lesion candidate area, and lesion information of a case on a lesion display unit.

Description

[0001] The present invention relates to an effective oral lesion onset detection method and system,

The present invention relates to a method for predicting or detecting an oral lesion, wherein a panoramic image of the oral cavity is set using a feature vector of an area case, etc., and then focused on a set search area, The present invention relates to a method for predicting the onset of oral lesions in which a lesion is searched in a panoramic image to perform a rapid and precise search for lesion detection.

In general, data on patients are collected in the search for lesions. X-ray, ultrasound, and radiation are used as techniques for the diagnosis, and various diagnoses and lesions can be detected therefrom.

In particular, the panoramic image of the oral image provides the image so that all the teeth can be seen on one image as the oral teeth spread sideways.

Thus, these panoramic images can be directly diagnosed by the doctors by the naked eye, such as the tooth condition, the dental condition, and the oral condition. In this way, panoramic images of the oral cavity can be diagnosed in real time by the doctors themselves.

However, in the case of the automatic lesion system, there are many lesion images to be diagnosed automatically, and since it is necessary to precisely contrast with the entire panorama image, it is necessary to cope with the problem because the surgeon may be slower than the doctor. In addition, if individual doctors depend on visual examination alone, diagnostic contents may be changed according to the subjective opinion of a doctor, so that it is difficult to provide a consistent treatment prescription.

Published Patent No. 10-2009-0118670 (Published Nov. 18, 2009) Open Patent No. 10-2013-0011823 (published on January 30, 2013) Open Patent No. 10-2014-0055250 (Published May 2014, 2014)

In order to solve the above-described problems, the present invention has an object to enable more accurate and precise lesion prediction by comparing a panoramic image of a patient's mouth with lesion case information in detecting and detecting an oral lesion.

It is another object of the present invention to provide a method and apparatus for performing a matching step with area case information on a panoramic image in which a tooth of a whole oral cavity spreads out before a lesion is searched to find a search area of a region- The panoramic image is automatically set so that accurate lesion search for the search area in which the subsequent lesion search is set is performed.

It is another object of the present invention to provide a method and apparatus for detecting a lesion using a lesion case only for a search region corresponding to an oral lesion such as a tooth among panoramic images of a patient.

It is yet another object of the present invention to make it easy to look at the lesion by making the detected lesion position appear and displayed in a predetermined region of interest (ROI).

According to an aspect of the present invention, there is provided a method for predicting the onset of oral lesions by a system for predicting the onset of oral lesions, the method comprising: searching for a lesion on a candidate image of a panoramic image, A search area setting step of setting a search area for searching for a search area; A lesion searching step of loading a lesion case image for lesion search from a data storage unit and detecting a lesion candidate region compared with the lesion case image among the search regions set in the search region setting step; And a lesion display step of displaying the lesion candidate information on the lesion candidate area, the lesion candidate information corresponding to the lesion candidate area, the lesion information of the lesion candidate area, and the lesion information of the case on the lesion display part, Wherein the image information displayed in the display step is composed of image information displayed in a region of interest (ROI) with respect to the lesion candidate region searched for in the lesion search step. to provide.

In a preferred embodiment of the present invention, a case information storing step of storing case information for discrimination in a data storage unit; And a discrimination target image information storing step of storing the discrimination target image data in a data storing unit, wherein the case information includes area case information for setting a search area for the panorama image of the patient, Wherein the region case information includes image data of a region case, a feature vector of a region case image, and region information of a search for a lesion search. Thereby providing an effective method of predicting the onset of oral lesions.

In a preferred embodiment of the present invention, the search region setting step includes: a case feature vector loading step of loading a feature vector of one or more domain case information from a data storage unit; A feature vector retrieving step of retrieving an area feature vector having a matching rate and a feature vector of a candidate image by comparing feature vectors of at least one area feature feature vector and a candidate image of the panoramic image; And a discrimination area setting step of setting a search area in the panorama candidate image so as to correspond to the area information of the search in the area case information retrieved in the feature vector retrieving step. ≪ / RTI >

In a preferred embodiment of the present invention, an optimal search area setting unit sets a search area for searching for a lesion in a candidate image of a panoramic image in comparison with area case information loaded from a data storage unit; A lesion search unit for loading a lesion case image for lesion search from a data storage unit and detecting a lesion candidate region compared with the lesion case image among search regions set in the optimum search region setting unit; The lesion candidate information, the lesion candidate information, and the lesion information of the case are displayed on the lesion display unit, and the displayed lesion candidate information is displayed on the lesion display unit, A lesion display unit for displaying a region of interest with image information displayed in a region of interest (ROI); And a data storage unit for storing the area case information, the lesion case information, the discrimination object information, and the information of the lesion candidate area.

In a preferred embodiment of the present invention, the optimal search area setting unit includes: a case feature vector loading module that loads feature vectors of one or more area case information from a data storage unit; A feature vector search module that compares at least one region feature feature vector with a feature vector of a candidate image of the panoramic image to search for a feature feature vector of the candidate image and an area feature vector having a high matching rate; And a discrimination area setting module for setting a search area in the panorama candidate image in correspondence with the area information of the search in the area case information searched by the feature vector search module. .

According to the present invention configured as described above, the present invention for solving the above-mentioned problems is characterized in that, in searching for and detecting an oral lesion, comparing a panoramic image of a patient's mouth with lesion case information, There is an effect that is possible.

The other effect of the present invention is that the panoramic image of the entire oral cavity spreading laterally is subjected to the matching step with the area case information prior to the lesion search to find the search area of the area case having the similar mouth cavity, The panoramic image is automatically set so that the subsequent search for the lesion can be performed accurately for the search area.

Further, another effect of the present invention is that a lesion is searched using a lesion case only for a search region corresponding to an oral lesion such as a tooth in a panoramic image of a patient, so that the lesion can be searched accurately and quickly.

Further, another effect of the present invention is that the detected lesion position can be easily viewed with respect to the lesion by displaying it in a predetermined region of interest (ROI).

FIG. 1 is a flowchart illustrating a method for predicting whether or not an oral lesion is present according to the present invention.
FIG. 2 is a diagram illustrating a detailed flow of a search region setting step in a method of predicting whether or not an oral lesion is present according to the present invention.
FIG. 3 is an explanatory diagram illustrating an application example in the search region setting step in the method of predicting whether or not to detect an oral lesion according to the present invention.
FIG. 4 is a diagram illustrating a detailed flow of a lesion search step in the method of predicting whether or not an oral lesion is present according to the present invention.
FIG. 5 is an explanatory diagram for explaining a comparison between lesion case image data and image data of a search target in a lesion search embodiment in the method for predicting whether or not to detect an oral lesion according to the present invention.
FIG. 6 is a detailed flowchart illustrating a method for displaying a lesion candidate region in a method for predicting whether or not a lesion is present according to the present invention.
FIG. 7 is an explanatory diagram showing an example of a lesion candidate region (ROI) detected in the method for predicting whether a lesion according to the present invention is predicted.
FIG. 8 is a flowchart illustrating an overall method of detecting whether or not an oral lesion is present according to the present invention.
FIG. 9 is an overall configuration diagram of a system for predicting whether or not an oral lesion according to the present invention is present.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

That is, as shown in FIGS. 1 to 9, the method for predicting the onset of oral lesions in a panorama according to the present invention and the system for predicting the onset of lesions of oral cavity according to the present invention, , The patient is compared with the oral image information of the newly photographed patient to detect the region of interest and the region of high matching rate so as to determine whether there is a region of interest similar to or similar to the lesion case, You can look at it.

Particularly, in the present invention, the image information photographed by the panoramic image is used. By comparing the analysis information such as the feature vector of the taken panoramic image and the feature vector of the case image, the region where the case and the matching ratio are high is referred to as a lesion candidate region (ROI) in the lesion candidate region, so that the lesion can be easily examined.

Hereinafter, an effective method for predicting the incidence of oral lesions in the panorama according to the present invention, and specific details of the system for predicting the incidence of oral lesions will be described with reference to the accompanying drawings.

The system for predicting lesion onset according to the present invention is provided with a member input by an administrator, a display which is a display part for viewing a lesion-detected image, and a configuration for lesion search. The panoramic image data stored in the panoramic image storage unit may be stored in the storage unit of the panoramic image storage unit through the input member managed by the manager, .

In addition, case information including a case image for image discrimination of whether the image data of the photographed patient corresponds to a known lesion is stored in the data storage unit 60 in an organized state by case.

That is, in the method for predicting the onset of oral lesions according to the present invention, a case information storing step (S10) for storing case information for discrimination is first performed in a data storing unit.

The case information includes area case information for setting a search area for a panorama image of a patient and lesion case information for lesion search for a search area, and the area case information includes image data of area cases A feature vector for the area case image, and area information of the search for the lesion search (e.g., coordinate values of the area from the reference point). Therefore, the region of the lesion is determined from the panoramic image of the patient using the case information, and the lesion portion is searched from the patient's panoramic image using the lesion case information.

 In addition, a panoramic image of a patient is provided in a separate storage device, or a patient image is provided from a panoramic image capturing device. In this case, the image data of a patient to be discriminated is stored in a data storage unit (S20).

Thus, the case information, which is basically a criterion for lesion discrimination, and the panoramic image information of the patient to be a lesion discrimination are all stored in the data storage unit 60. In other words, the data storage unit 60 stores data such as area case information, lesion case information, information on a discrimination target, information on a lesion candidate region, and the like.

Accordingly, the data storage unit 60 is provided with a configuration in which data for management, control configuration for lesion search, and other data for driving are stored. And a case information storage unit 61 in which case information is stored. The case information storage unit 61 includes area case information for setting a search area in the panorama image of the patient, sets the search area in the patient's panoramic image, and then searches for the lesion by focusing on the search area Case information including case information is stored.

And a discrimination image storage unit 62 in which panorama image information of a patient to be discriminated is inputted and stored is provided. And then the configuration for the search is performed by loading the case image data of the case information and the image information of the patient.

As shown in the accompanying drawings, the panoramic image information is a plan view image of the entire oral teeth spread in the lateral direction. In other words, the structure of the oral cavity is in the form of a U-shape when viewed from the horizontal plane, but the image is taken with the oral teeth spreading sideways as a whole using X-ray or the like. Therefore, it is a video that makes it easy to see many teeth and oral condition at a glance. In the present invention, a lesion is searched for a panoramic image of the oral cavity.

In the data storage unit 60, the data of the process of searching for a lesion on the patient's panoramic image using the lesion case information is stored in the search temporary storage unit 63, and the search process is performed. The process is fast and relieves the system of data overload.

The data of the lesion region detected in the search process is stored in the lesion detection information storage unit 64, and such data is displayed on the display of the lesion display unit 50. [

As described above, the data storage unit 60 stores the case information and the panoramic image information to be discriminated by the patient. Then, the patient's image is analyzed through a lesion search process to determine whether there is a portion coinciding with or similar to the image of the case, .

The configuration and processing steps for this are described below.

Accordingly, in the oral lesion disease predicting system according to the present invention, an optimal search area setting unit for setting a search area for searching for a lesion in a candidate image of a panoramic image in comparison with the area case information loaded from the data storage unit 60 30).

In addition, in the method for predicting the onset of oral lesions by the system for predicting the onset of oral lesions, the optimal search area setting unit 30 as shown in FIG. And a search area setting step (S30) for setting a search area for searching for a lesion in the candidate image.

According to the optimal search area setting unit 30 and the search area setting step S30, the search area is set using the area case information for the panorama image of the patient to be identified, and then the lesion case information is used for the search area A lesion search is performed.

For this purpose, first, an area for searching is set in the panoramic image of the patient. As shown in FIG. 3, a panoramic image of the oral cavity is searched for a case information having a high matching rate through comparison with a regional case image. By applying the search region information of the case to the patient's panoramic image, .

The detailed configuration will be described with reference to FIG. 2, FIG. 3, and the like.

The optimal search area setting unit 30 includes a case feature vector loading module 31 that loads feature vectors of one or more domain case information from the data storage unit 60.

The searching region setting step S30 may include a case feature vector loading step S31 for loading a feature vector of one or more domain case information from the data storage unit 60 by the case feature vector loading module 31, .

Such area case information includes a plurality of area cases as shown in FIG. 3, and FIG. 3 shows four area case images. Each of these area case information includes feature vectors and setting information of the search area.

Next, the optimal search area setting unit 30 compares one or more area case feature vectors with the feature vectors of the candidate images of the panoramic image to obtain a feature feature vector of the candidate image and a region case feature vector having a high matching rate And a feature vector search module 32 for searching for the feature vector.

In addition, the search region setting step S30 may include a feature vector search module 32 for comparing the feature vector of one or more region feature vectors with the feature vector of the candidate image of the panoramic image, A feature vector search step (S32) for searching the high region feature vector is performed.

That is, the feature vectors of the plurality of region case information and the feature vector of the patient's panorama image are compared and discriminated, and region case information having a high matching rate is selected.

The characteristic vector in the image analysis can be applied to the analysis using the feature vector. For example, the feature vector in the oral image includes the tooth structure, the tooth position, the interval information between the maxilla and mandible, For example, a black-and-white image in these images is transformed so as to be a numerical value or a comparison with respect to a brightness distribution, a position value of a bright point, a degree of brightness change, an image boundary, It is useful for comparing feature elements. By applying a feature vector that is generally used, an area having a high matching ratio with a case image is searched and set.

The optimal search area setting unit 30 sets a search area in the panorama candidate image in correspondence with the search area information in the area case information searched in the feature vector search step S32 and the feature vector search module 32, And a discrimination area setting module 33 for discriminating the discrimination area.

In addition, the search area setting step S30 corresponds to the search area information of the area case information searched by the feature vector search step S32 and the feature vector search module 32 by the determination area setting module 33 A discrimination area setting step (S33) for setting a search area in the panorama candidate image is performed.

Next, the lesion detection process will be described.

In the oral cavity lesion predicting system, the lesion case image for lesion search is loaded from the data storage unit 60, and the lesion case image is compared with the lesion case image among the search regions set in the optimum search region setting unit 30. [ And a lesion search unit 40 for detecting a lesion candidate region.

In the method for predicting the invention of oral cavity lesion, a lesion case image for lesion search is loaded from the data storage unit 60 by the lesion search unit 40, and the lesion case image is set in the search area setting step S30 A lesion searching step (S40) for detecting a lesion candidate region in contrast to the lesion case image is performed in the search region.

This lesion search is to perform a lesion search for a region of interest after setting the region of interest in the panoramic image of the patient.

The detailed configuration of the lesion search unit 40 and the lesion search step S40 will be described with reference to FIG. 4 and FIG. 5 and the like.

The lesion search unit 40 includes a case information loading module 41 for loading lesion case information for lesion detection for a panorama candidate image to be discriminated from the data storage unit 60.

In addition, the lesion searching step S40 may include a case information loading step of loading lesion case information for lesion search for the panorama candidate image to be discriminated by the case information loading module 41 from the data storage unit 60 (S41).

Such lesion cases are for lesion detection on the oral panoramic image, and the lesion case information may include feature vector information, luminance of the image, noise, noise, brightness, position information on the lesion of the entire oral image, etc. .

Next, the lesion search unit 40 includes a lesion region search module 42 for comparing lesion regions of the lesion case image and the lesion region with respect to the lesion case image by comparing and discriminating images of the lesion case image and the candidate region.

In addition, the lesion search step S40 may include a lesion area searching module 42 for comparing the lesion case image and the image of the search region of the candidate image to search for a lesion area contrasted with the lesion case image, Step S42 is performed.

That is, in the process of the preceding search area setting step (S30), a search area is set for the portion of the oral cavity in the patient's panoramic image, and the feature vector (tooth structure, teeth Position, distance between the maxilla and mandible, dental arrangement, etc.) and information on luminance, noise, noise, brightness, etc. by each tooth are searched for a portion having lesion characteristics.

Also, in the case of a lesion that appears specifically on the site in terms of coordinates, the focus may be on lesions that may be seen on the coordinates.

In this way, the lesion is searched and the searched region is defined as the searched region.

That is, the lesion search unit 40 sets the lesion region as a lesion candidate region by searching the lesion region for each lesion case by the lesion region search module 42, and stores the lesion search information in the database 60 And a lesion candidate region processing module 43.

In the lesion searching step S40, the lesion candidate region processing module 43 sets the lesion region as a lesion candidate region by searching for the lesion for the individual lesion case in the lesion region searching step S42, A lesion candidate region processing step S43 for storing the search information in the database 60 is performed.

In other words, the region to be searched can be set as a region of interest (ROI) as a lesion candidate region, so that it can be easily looked at at a later stage.

Next, in order to process the searched lesion area after the lesion search using the lesion case as described above, image information on the lesion candidate area searched by the lesion search unit 40 as shown in Figs. 6 and 7, The lesion candidate information area, the lesion candidate information area, and the lesion information of the case are displayed on the lesion display unit 50, and the displayed image information is displayed as the image information displayed in the ROI on the detected lesion candidate area And a lesion display unit 50.

A lesion display step of displaying the lesion candidate area detected by the lesion search step (S40) on the lesion display part (50) with image information, corresponding case image information, information data of the lesion candidate area, (S50).

As shown in FIG. 7, the image information displayed in the lesion display step S50 may be displayed as a region of interest (ROI) for the lesion candidate region searched for in the lesion search step S40 to view a lesion panorama image (ROI) of the portion of the lesion that can be regarded as a lesion, so that the prediction of lesion onset can be easily achieved.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. The technical idea of the present invention should not be construed as being limited.

30: Optimum search area setting unit
40:
50:
60: Data storage unit

Claims (5)

A method for predicting the onset of oral lesions by a system for predicting the onset of oral lesions,
A search area setting step of setting a search area for searching for a lesion in a candidate image of the panoramic image in comparison with the area case information loaded from the data storage part;
A lesion searching step of loading a lesion case image for lesion search from a data storage unit and detecting a lesion candidate region compared with the lesion case image among the search regions set in the search region setting step; And
A lesion display step of displaying, on a lesion display unit, the image information of the lesion candidate region searched for by the lesion search step, the relevant case image information, the information data of the lesion candidate region, and the lesion information of the case;
Lt; / RTI >
Wherein the image information displayed in the lesion display step is composed of image information displayed in an ROI on a lesion candidate region searched for in the lesion search step. Way.
The method according to claim 1,
A case information storing step of storing case information for discrimination in a data storing unit; And
And storing the image data to be discriminated in a data storage unit,
The above-
Area case information for setting a search area for a panorama image of a patient,
A lesion case information for lesion search for a search region,
Wherein the area case information includes image data of the area case, feature vector of the area case image, and area information of the search for the lesion search.
The method according to claim 1,
Wherein the search area setting step comprises:
Loading a feature vector of one or more region case information from a data storage unit;
A feature vector retrieving step of retrieving an area feature vector having a matching rate and a feature vector of a candidate image by comparing feature vectors of at least one area feature feature vector and a candidate image of the panoramic image; And
A discrimination area setting step of setting a search area in the panorama candidate image so as to correspond to the area information of the search in the area case information retrieved in the feature vector retrieving step;
The method of claim 1, further comprising:
An optimal search area setting unit for setting a search area for searching for a lesion in a candidate image of the panoramic image in comparison with the area case information loaded from the data storage unit;
A lesion search unit for loading a lesion case image for lesion search from a data storage unit and detecting a lesion candidate region compared with the lesion case image among search regions set in the optimum search region setting unit;
The lesion candidate information, the lesion candidate information, and the lesion information of the case are displayed on the lesion display unit, and the displayed lesion candidate information is displayed on the lesion display unit, A lesion display unit for displaying a region of interest with image information displayed in a region of interest (ROI); And
A lesion candidate information area, a lesion case information, a discrimination target information, and a lesion candidate area, and a data storage unit for storing the lesion candidate area information.
5. The method of claim 4,
Wherein the optimum search area setting unit comprises:
A case feature vector loading module loading a feature vector of one or more region case information from a data store;
A feature vector search module that compares at least one region feature feature vector with a feature vector of a candidate image of the panoramic image to search for a feature feature vector of the candidate image and an area feature vector having a high matching rate; And
And a discrimination area setting module for setting a search area in the panorama candidate image in correspondence with the area information of the search in the area case information searched by the feature vector search module.

Images