A method and system for detecting oral plaque markers

Abstract

The application relates to the technical field of image analysis, in particular to a dental plaque marking detection method and system, which comprises the following steps: acquiring tooth surface images of a patient to be detected under different preset wavelengths, and identifying a target tooth region from each tooth surface image; determining a plaque feature performance value and a plaque possible width corresponding to each tooth pixel point; taking the tooth pixel point as an endpoint, drawing a target ray with an extending direction being a vertical upward direction, and forming a dental crown intersection segment by intersecting the target ray with a dental crown region; cutting a reference pixel segment from the target ray; determining a plaque possible index corresponding to each tooth pixel point; fusing all the tooth surface images based on the plaque possible index, and performing dental plaque marking detection based on the obtained fused image. The application realizes marking detection of dental plaque by analyzing tooth surface images under different preset wavelengths, and improves the accuracy of dental plaque marking detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image analysis, in particular to a dental plaque marking detection method and system. BACKGROUND

[0002] With the development of science and technology, image analysis is applied more and more widely, for example, it can be applied to dental plaque marking detection. Dental plaque on the surface of teeth is generated by the combination of bacteria and protein in food residues and saliva. Because saliva is easy to form longitudinal shear force when chewing teeth, bacteria are more likely to root in the position of longitudinal low shear force. Therefore, it is very important to detect dental plaque marking on lower teeth. At present, when detecting dental plaque marking, the method commonly used is to detect dental plaque marking by the tooth image collected by a hyperspectral camera at a single detection wavelength.

[0003] However, when detecting lower dental plaque marking by the tooth image collected by a hyperspectral camera at a single detection wavelength, the following technical problems often exist:

[0004] Because the characteristics of dental plaque on the images collected by different hyperspectral cameras at different detection wavelengths are often different, when detecting lower dental plaque marking by the tooth image collected by a hyperspectral camera at a single detection wavelength, the characteristics of dental plaque analyzed may be relatively single, thereby leading to poor accuracy of dental plaque marking detection. SUMMARY

[0005] In order to solve the technical problem of poor accuracy of dental plaque marking detection, the present application provides a dental plaque marking detection method and system.

[0006] In a first aspect, the present application provides a dental plaque marking detection method, which comprises:

[0007] Obtaining tooth surface images of a patient to be detected at different preset wavelengths, and identifying a target tooth region from each tooth surface image, wherein the tooth surface image is a hyperspectral image, one target tooth region represents one lower tooth, and the target tooth region includes a tooth crown region;

[0008] Determining a plaque characteristic performance value corresponding to each tooth pixel point according to a gray value corresponding to each tooth pixel point in each target tooth region and a preset wavelength to which the gray value belongs;

[0009] Determining a plaque possible width corresponding to each tooth pixel point according to the plaque characteristic performance values corresponding to all tooth pixel points in a preset horizontal window corresponding to each tooth pixel point;

[0010] Taking the tooth pixel point as an end point, a target ray with an extending direction being a vertical upward direction is drawn, and an intersection of the target ray and a crown region is a crown intersection line segment corresponding to the tooth pixel point;

[0011] A reference pixel segment is cut from the target ray corresponding to each tooth pixel point according to the crown intersection line segment corresponding to each tooth pixel point;

[0012] A plaque possible index corresponding to each tooth pixel point is determined according to a difference between the plaque possible widths corresponding to the tooth pixel point in the reference pixel segment corresponding to each tooth pixel point.

[0013] Based on the plaque possible index, all tooth surface images are fused, and plaque labeling detection is performed based on the obtained fused image.

[0014] In a possible implementation manner of the first aspect, the determining of the plaque feature performance value corresponding to each tooth pixel point according to the gray value corresponding to each tooth pixel point in each target tooth region and the preset wavelength to which the tooth pixel point belongs includes:

[0015] An arbitrary tooth pixel point is determined as a marker pixel point, if the preset wavelength to which the marker pixel point belongs is in a first preset wavelength range, a gray value corresponding to the marker pixel point is determined as the plaque feature performance value corresponding to the marker pixel point.

[0016] If the preset wavelength to which the marker pixel point belongs is in a second preset wavelength range, a difference between a maximum value of a preset image gray range and the gray value corresponding to the marker pixel point is determined as the plaque feature performance value corresponding to the marker pixel point.

[0017] In a possible implementation manner of the first aspect, the determining of the plaque possible width corresponding to each tooth pixel point according to the plaque feature performance values corresponding to all tooth pixel points in the preset horizontal window corresponding to each tooth pixel point includes:

[0018] A mean value of the plaque feature performance values corresponding to all tooth pixel points in all target tooth regions in each tooth surface image is determined as an overall feature performance value corresponding to each tooth surface image.

[0019] The plaque possible width corresponding to each tooth pixel point is determined according to a difference between the plaque feature performance values corresponding to all tooth pixel points in the preset horizontal window corresponding to each tooth pixel point and the overall feature performance value corresponding to the tooth surface image to which the tooth pixel point belongs.

[0020] In a possible implementation manner of the first aspect, the difference between the plaque feature value corresponding to each tooth pixel point and the overall feature value corresponding to the tooth surface image to which the tooth pixel point belongs is determined according to the plaque feature value corresponding to each tooth pixel point in the preset horizontal window corresponding to the tooth pixel point, and the plaque possible width corresponding to each tooth pixel point is determined according to the difference between the plaque feature value corresponding to each tooth pixel point and the overall feature value corresponding to the tooth surface image to which the tooth pixel point belongs.

[0021] An arbitrary tooth pixel point is determined as a marker pixel point, and the overall feature value corresponding to the tooth surface image to which the marker pixel point belongs is determined as a marker overall value.

[0022] The mean value of the difference between the marker overall value and the plaque feature value corresponding to each tooth pixel point in the preset horizontal window corresponding to the marker pixel point is normalized to obtain the plaque possible width corresponding to the marker pixel point.

[0023] In a possible implementation manner of the first aspect, the reference pixel segment is cut from the target ray corresponding to each tooth pixel point according to the intersection line segment of the dental crown corresponding to each tooth pixel point.

[0024] An arbitrary tooth pixel point is determined as a marker pixel point, and the tooth pixel point farthest from the marker pixel point is selected from the intersection line segment of the dental crown corresponding to the marker pixel point as a reference pixel point corresponding to the marker pixel point.

[0025] The marker pixel point and the reference pixel point corresponding to the marker pixel point are connected to obtain the reference pixel segment corresponding to the marker pixel point.

[0026] In a possible implementation manner of the first aspect, the plaque possible index corresponding to each tooth pixel point is determined according to the difference between the plaque possible width corresponding to each tooth pixel point in the reference pixel segment corresponding to each tooth pixel point.

[0027] The mean value of the difference between the plaque possible width corresponding to each tooth pixel point and the plaque possible width corresponding to all tooth pixel points in the reference pixel segment corresponding to each tooth pixel point is normalized to obtain the plaque initial suspected factor corresponding to each tooth pixel point.

[0028] The plaque reference suspected factor corresponding to each tooth pixel point is determined according to the difference between the plaque possible width corresponding to adjacent tooth pixel points in the reference pixel segment corresponding to each tooth pixel point.

[0029] The plaque possible index corresponding to each tooth pixel point is determined according to the plaque possible width, the plaque initial suspected factor and the reference suspected factor corresponding to each tooth pixel point.

[0030] In a possible implementation manner of the first aspect, the method further includes: determining, according to the difference between the possible widths of plaque corresponding to adjacent tooth pixel points in the reference pixel segment corresponding to each tooth pixel point, a plaque reference suspicious factor corresponding to each tooth pixel point.

[0031] In a possible implementation manner of the first aspect, the method further includes: sorting, in a top-down order, the pixel points in the reference pixel segment corresponding to each tooth pixel point to obtain a reference pixel sequence corresponding to each tooth pixel point.

[0032] In a possible implementation manner of the first aspect, the method further includes: normalizing the mean of the difference between the possible widths of plaque corresponding to adjacent tooth pixel points in the reference pixel sequence corresponding to each tooth pixel point to obtain the plaque reference suspicious factor corresponding to each tooth pixel point.

[0033] In a possible implementation manner of the first aspect, the method further includes: fusing all the dental surface images based on the plaque possibility index.

[0034] In a possible implementation manner of the first aspect, the method further includes: fusing the tooth pixel points in all the dental surface images according to the plaque feature performance value and the plaque possibility index of the tooth pixel points at the same position in all the dental surface images to obtain a fusion image composed of all the fused tooth pixel points.

[0035] In a possible implementation manner of the first aspect, the method further includes: performing dental plaque marking detection based on the obtained fusion image.

[0036] In a possible implementation manner of the first aspect, the method further includes: performing segmentation on the obtained fusion image by using the maximum inter-class variance method to obtain two sub-regions, and marking the sub-region with a smaller gray value in the two sub-regions as a dental plaque region.

[0037] In a possible implementation manner of the first aspect, the method further includes: performing dental plaque marking detection based on the obtained fusion image.

[0038] The acquisition and identification module is configured to acquire dental surface images of a to-be-detected patient under different preset wavelengths, and identify a target tooth region from each dental surface image.

[0039] The plaque feature performance value determination module is configured to determine a plaque feature performance value corresponding to each tooth pixel point according to a gray value of each tooth pixel point and a preset wavelength to which the tooth pixel point belongs.

[0040] The plaque possibility width determination module is configured to determine a plaque possibility width corresponding to each tooth pixel point according to the plaque feature performance values of all tooth pixel points in a preset horizontal window corresponding to each tooth pixel point.

[0041] A ray and line segment construction module is configured to take the tooth pixel point as an end point, extend a target ray in a vertical upward direction, and take the intersection of the target ray and the crown region as a tooth pixel point corresponding crown intersection line segment.

[0042] A pixel segment segmentation module is configured to cut a reference pixel segment from the target ray corresponding to each tooth pixel point according to the tooth pixel point corresponding crown intersection line segment corresponding to each tooth pixel point.

[0043] A plaque possibility index determination module is configured to determine the plaque possibility index corresponding to each tooth pixel point according to the difference between the plaque possibility widths corresponding to each tooth pixel point in the reference pixel segment corresponding to each tooth pixel point.

[0044] A fusion and detection module is configured to fuse all tooth surface images based on the plaque possibility index, and detect the dental plaque label based on the obtained fused image.

[0045] In a third aspect, a server is provided, including a memory and a processor. The memory is configured to store executable program code, and the processor is configured to call and run the executable program code from the memory, so that the device executes the method in the first aspect or any possible implementation manner of the first aspect.

[0046] In a fourth aspect, a computer program product is provided, which includes computer program code. When the computer program code is run on a computer, the computer executes the method in the first aspect or any possible implementation manner of the first aspect.

[0047] In a fifth aspect, a computer readable storage medium is provided, which stores computer program code. When the computer program code is run on a computer, the computer executes the method in the first aspect or any possible implementation manner of the first aspect.

[0048] The present application has the following beneficial effects:

[0049] The present invention provides an oral plaque marking detection method that achieves plaque marking detection by analyzing tooth surface images at different preset wavelengths, thereby resolving the technical problem of poor accuracy in plaque marking detection and improving the accuracy of plaque marking detection. Specifically, the present invention comprehensively analyzes the characteristic manifestations of dental plaque in tooth surface images at different preset wavelengths, specifically quantifies multiple indicators related to plaque manifestation characteristics, such as plaque characteristic manifestation values, plaque possible width, and plaque possible indicators, thereby achieving fusion of all tooth surface images, and achieving plaque marking detection based on the obtained fused image, thereby improving the accuracy of plaque marking detection on lower teeth, thereby improving the accuracy of plaque marking detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 This is a flow chart of a method for detecting oral dental plaque markers according to the present invention;

[0052] Figure 2 Schematic diagram of the structure of an oral dental plaque marking detection system of the present invention;

[0053] Figure 3 The figure is a structural diagram of a computer device of the present invention. DETAILED DESCRIPTION

[0054] To further illustrate the technical means and effects employed by the present invention to achieve its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementations, structures, features, and effects of the technical solutions proposed by the present invention. In the following description, references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0055] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0056] Dental plaque is a layer of biofilm on the surface of teeth, mainly composed of bacteria and food residues, which is usually difficult to see directly with the naked eye and needs special staining agents or imaging techniques to visualize. Common staining agents include erythrosin, basic fuchsin, sodium fluorescein, etc., which can make dental plaque color. However, some patients may have poor staining effect due to licking, swallowing, etc., and there may be safety risks. Therefore, non-staining agent staining, dental plaque marking scheme through multispectral imaging technology is gradually popular.

[0057] The surface of dental plaque and enamel is relatively smooth, which is difficult to separate by traditional image threshold segmentation algorithm. Dental plaque contains various components, and the current fluorescence detection method based on threshold segmentation uses blue-green fluorescence (405 nm) to irradiate the surface of the tooth, uses the strong absorption of porphyrin components in dental plaque to 600-700 nm, and only receives the red fluorescence signal image through the 610 nm long-pass filter, and carries out threshold segmentation and plaque recognition on the image. However, it ignores the characteristics of dental plaque that can be reflected under other detection wavelengths, lacks a method to scientifically integrate image information under other wavelengths, and makes the final plaque marking segmentation effect poor and the accuracy low.

[0058] The present application comprehensively analyzes the characteristic performance of dental plaque in the tooth surface image under different preset wavelengths, specifically quantifies a plurality of plaque performance characteristic-related indexes such as plaque characteristic performance value, plaque possible width and plaque possible index, etc., thereby realizing the fusion of all tooth surface images, and based on the obtained fused image, realizing the marking detection of dental plaque, improving the accuracy of the marking detection of dental plaque of the lower teeth, thereby improving the accuracy of the marking detection of dental plaque.

[0059] Reference Figure 1 , shows the flow of some embodiments of the oral dental plaque marking detection method of the present application. The oral dental plaque marking detection method comprises the following steps:

[0060] Step S1, acquiring the tooth surface images of the patient to be detected under different preset wavelengths, and identifying the target tooth region from each tooth surface image.

[0061] The patient to be detected can be a patient to be detected for dental plaque marker detection. The preset wavelength can be a preset detection wavelength when an image is captured by the hyperspectral camera. The preset wavelength can be in a first preset wavelength range and a second preset wavelength range. The first preset wavelength range can represent a lower detection wavelength range of the hyperspectral image, which can be set to 440-520 nm. The second preset wavelength range can represent a higher detection wavelength range of the hyperspectral image, which can be set to 620-670 nm. The difference between each adjacent preset wavelength can be 10 nm. The tooth surface image at the preset wavelength can be a hyperspectral image of the tooth surface of the patient to be detected captured by the hyperspectral camera when the detection wavelength is equal to the preset wavelength. One target tooth region can represent one lower tooth of the patient to be detected. The target tooth region can include a tooth crown region, a tooth neck region, and a tooth root region.

[0062] As an example, the step can include the following steps:

[0063] First, acquire the tooth surface images of the patient to be detected at different preset wavelengths.

[0064] For example, a semiconductor laser or LED with a wavelength range of 400-410 nm and a center wavelength of 405 nm can be used. The light beam is focused and collimated to a parallel light with a diameter of 0.2 mm, and is obliquely incident on the tooth surface at an angle of 45°. At this time, the hyperspectral camera is adjusted by adjusting the filter wheel to allow different wavelengths to pass through, and the obtained image is the tooth surface image. For example, if the preset wavelength is 500 nm, when the tooth surface image at 500 nm is acquired, the wavelength range allowed to pass through by the hyperspectral camera can be 495-505 nm, with a center wavelength of 500 nm. In the tooth surface image at 500 nm, each pixel point belongs to a preset wavelength of 500 nm.

[0065] It should be noted that the patient first performs mouth washing, then uses a dental clamp to fix the teeth, and then uses the hyperspectral camera to acquire the multi-wavelength image. During the process of capturing the hyperspectral image, the focal length and imaging conditions are first adjusted, and the patient's posture and camera angle are kept unchanged.

[0066] Second, identify the tooth region representing each lower tooth of the patient to be detected from each tooth surface image by a neural network method or a threshold segmentation method, and mark it as a target tooth region.

[0067] It should be noted that since there is a clear contour between the teeth and the gums, the neural network method or the threshold segmentation method can be used for tooth contour extraction. Each tooth contour often represents a tooth, so the tooth region representing each upper tooth and the tooth region representing each lower tooth can be obtained, and the tooth region of each lower tooth can be marked as a target tooth region.

[0068] Step S2, according to the gray value corresponding to each tooth pixel point in each target tooth area and the preset wavelength to which it belongs, determine the plaque feature performance value corresponding to each tooth pixel point.

[0069] As an example, the present step can include the following steps:

[0070] First, determine any tooth pixel point as a marker pixel point, if the preset wavelength to which the marker pixel point belongs is in the first preset wavelength range, then determine the gray value corresponding to the marker pixel point as the plaque feature performance value corresponding to the marker pixel point. The formula can be:

[0071] ;

[0072] Wherein, q is the plaque feature performance value corresponding to the marker pixel point. is the gray value corresponding to the marker pixel point.

[0073] Second, if the preset wavelength to which the marker pixel point belongs is in the second preset wavelength range, then determine the difference between the maximum value of the preset image gray range and the gray value corresponding to the marker pixel point as the plaque feature performance value corresponding to the marker pixel point. The formula can be:

[0074] ;

[0075] Wherein, q is the plaque feature performance value corresponding to the marker pixel point. The preset image gray range is the gray range of the image set in advance, which is often [0, 255]. A is the maximum value of the preset image gray range, which is equal to 255. is the gray value corresponding to the marker pixel point.

[0076] It should be noted that in the first preset wavelength range representing the lower detection wavelength interval of the hyperspectral image, such as the 440-520nm interval, because the normal tooth plaque tissue reflects efficiently and the tooth plaque tissue absorbs efficiently, the gray feature of the tooth plaque is significantly smaller than that of the normal tissue. On the contrary, in the second preset wavelength range representing the higher detection wavelength interval, such as the 620-670nm interval, the tooth plaque emits red light in this detection interval, and the normal tissue shows almost none, so the gray feature of the tooth plaque is significantly larger than that of the normal tissue. Therefore, the greater the plaque feature performance value corresponding to the tooth pixel point, the more likely it is that the tooth pixel point is a tooth plaque pixel point.

[0077] Step S3, determining the possible width of the plaque corresponding to each tooth pixel point according to the plaque characteristic expression values ​​corresponding to all tooth pixels points within the preset horizontal window corresponding to each tooth pixel point.

[0078] The preset horizontal window may be a rectangular window with a width of 1 and a size of 1 × 7. The tooth pixel point may be located at the center of its corresponding preset horizontal window.

[0079] It's important to note that dental plaque typically exhibits certain morphological characteristics. For example, when a tooth's surface is covered with plaque, it often forms patches at the root, while at the cusp, due to long-term wear, plaque often appears in longitudinal cracks. Therefore, we can analyze the plaque width corresponding to a tooth pixel. A larger value indicates that the tooth pixel is more likely to be a plaque pixel, and the plaque surrounding that tooth pixel is more likely to form patches.

[0080] As an example, this step may include the following steps:

[0081] In the first step, the average of the plaque characteristic expression values ​​corresponding to all tooth pixels in all target tooth areas in each tooth surface image is determined as the overall characteristic expression value corresponding to each tooth surface image.

[0082] The second step is to determine the possible plaque width corresponding to each tooth pixel point based on the difference between the plaque feature expression values ​​corresponding to all tooth pixels within the preset horizontal window corresponding to each tooth pixel point and the overall feature expression value corresponding to the tooth surface image to which it belongs, which may include the following sub-steps:

[0083] In the first sub-step, any tooth pixel is determined as a marked pixel, and the overall feature expression value corresponding to the tooth surface image to which the marked pixel belongs is determined as the marked overall expression value.

[0084] In the second sub-step, the average of the differences between the overall performance value of the mark and the plaque characteristic performance values ​​corresponding to each tooth pixel point in the preset horizontal window corresponding to the mark pixel point is normalized to obtain the possible width of the plaque corresponding to the mark pixel point.

[0085] For example, the formula for determining the possible width of plaque corresponding to a tooth pixel can be:

[0086] ;

[0087] in, It is i The possible width of plaque corresponding to each tooth pixel. i is the serial number of the tooth pixel. is the normalization function. It is i The number of tooth pixels in the preset horizontal window corresponds to the number of tooth pixels. j It is i The sequence number of the tooth pixel point in the preset horizontal window corresponding to each tooth pixel point. It is i The overall feature expression value of the tooth surface image corresponding to the tooth pixel point, that is, i The mean of the plaque characteristic values ​​corresponding to all tooth pixels in the tooth surface image to which the tooth pixel belongs. It is i In the preset horizontal window corresponding to the tooth pixel point, the j The plaque characteristic value corresponding to each tooth pixel.

[0088] It should be noted that when The larger the j The lower the plaque characteristic performance value corresponding to each tooth pixel is than the overall characteristic performance value representing the overall plaque characteristic performance of the patient's teeth, the higher the plaque characteristic performance value is. j The more tooth pixels there are, the more likely they are to be dental plaque pixels. The larger the i The more plaque pixels there are in the preset horizontal window corresponding to the first tooth pixel, the more likely it is that there are more dental plaque pixels. i The more likely the pixel point of the tooth is to be a dental plaque pixel, and the i The greater the plaque level around each tooth pixel, the greater the plaque level may be.

[0089] In step S4, a target ray is drawn with the tooth pixel point as an endpoint and the extension direction is vertically upward. The intersection of the target ray and the tooth crown area forms a tooth crown intersection line segment corresponding to the tooth pixel point.

[0090] The target ray may be a ray with the tooth pixel point as an end point and extending in a vertically upward direction.

[0091] It should be noted that the tooth crown is often located at the upper part of the tooth. Therefore, a ray with the tooth pixel as the endpoint and extending in the vertical upward direction often intersects the tooth area.

[0092] As an example, any tooth pixel point can be determined as a marked pixel point, and a ray with the marked pixel point as the endpoint and the extension direction in the vertical upward direction is drawn, which is recorded as the target ray corresponding to the marked pixel point, and the intersection of the target ray corresponding to the marked pixel point and the crown area included in the target tooth area to which the marked pixel point belongs is recorded as the crown intersection segment corresponding to the marked pixel point.

[0093] Step S5: according to the tooth crown intersection segment corresponding to each tooth pixel point, a reference pixel segment is cut out from the target ray corresponding to each tooth pixel point.

[0094] As an example, this step may include the following steps:

[0095] In the first step, any tooth pixel is determined as a marker pixel, and the tooth pixel farthest from the marker pixel is selected from the tooth crown intersection segment corresponding to the marker pixel as the reference pixel corresponding to the marker pixel.

[0096] The second step is to connect the marked pixel point and its corresponding reference pixel point to obtain the reference pixel segment corresponding to the marked pixel point.

[0097] It should be noted that the tooth pixel point on the reference pixel segment corresponding to the marked pixel point is often the tooth pixel point above the marked pixel point.

[0098] Step S6, determining a plaque possible index corresponding to each tooth pixel point according to the difference between possible plaque widths corresponding to the tooth pixel point in the reference pixel segment corresponding to each tooth pixel point.

[0099] It should be noted that dental plaque on the surface of teeth is formed by bacteria combining with proteins in food residues and saliva. Saliva easily forms longitudinal shear forces when chewing, and bacteria are more likely to take root in locations with low longitudinal shear forces. Furthermore, because tooth enamel is composed of vertically arranged enamel rods, the longitudinal shear forces between the rods are much smaller than the shear forces on the rods, resulting in a distinct striped distribution. However, the shear forces and wear at the root are usually much smaller than at the tip. The closer the plaque is to the root, the thicker it is, so the width of the longitudinal striped texture gradually widens, and may even form a patchy distribution.

[0100] As an example, this step may include the following steps:

[0101] In the first step, the average of the difference between the plaque possible width corresponding to each tooth pixel and the plaque possible width corresponding to all tooth pixels in its corresponding reference pixel segment is normalized to obtain the initial plaque suspicion factor corresponding to each tooth pixel.

[0102] For example, the formula for determining the initial plaque suspicion factor corresponding to the tooth pixel point can be:

[0103] ;

[0104] in, It is i The initial plaque suspicion factor corresponding to each tooth pixel. i is the serial number of the tooth pixel. is a normalization function. is the number of tooth pixel points in the reference pixel segment corresponding to the i th tooth pixel point. a is the sequence number of tooth pixel points in the reference pixel segment corresponding to the i th tooth pixel point. is the possible plaque width corresponding to the i th tooth pixel point. is the possible plaque width corresponding to the i th tooth pixel point in the reference pixel segment corresponding to the a th tooth pixel point.

[0105] It should be noted that the greater the , the more likely the possible plaque width corresponding to the i th tooth pixel point is higher than the possible plaque width corresponding to the tooth pixel points in the reference pixel segment corresponding to the i th tooth pixel point, and the more likely the th tooth pixel point is a plaque pixel point.

[0106] The second step of determining the plaque reference suspicion factor corresponding to each tooth pixel point according to the difference between the possible plaque widths corresponding to adjacent tooth pixel points in the reference pixel segment corresponding to each tooth pixel point can include the following sub-steps:

[0107] The first sub-step sorts the pixel points in the reference pixel segment corresponding to each tooth pixel point in order from bottom to top to obtain the reference pixel sequence corresponding to each tooth pixel point.

[0108] Among them, the tooth pixel point with a larger sequence number in the reference pixel sequence is more likely to be closer to the crown area and farther away from the root area.

[0109] The second sub-step normalizes the mean of the difference between the possible plaque widths corresponding to all adjacent tooth pixel points in the reference pixel sequence corresponding to each tooth pixel point to obtain the plaque reference suspicion factor corresponding to each tooth pixel point.

[0110] For example, the formula for determining the plaque reference suspicion factor corresponding to the tooth pixel point can be:

[0111] ;

[0112] Among them, is the plaque reference suspicion factor corresponding to the i th tooth pixel point. i is the sequence number of the tooth pixel point. is a normalization function. is the possible plaque width corresponding to the iThe number of tooth pixels in the reference pixel sequence corresponding to the tooth pixel. b It is i The sequence number of the tooth pixel in the reference pixel sequence corresponding to the tooth pixel. It is i In the reference pixel sequence corresponding to the tooth pixel point, the b The possible width of plaque corresponding to each tooth pixel. It is i In the reference pixel sequence corresponding to the tooth pixel point, the b +The possible width of plaque corresponding to 1 tooth pixel.

[0113] It should be noted that when The larger the i The plaque widths corresponding to the adjacent tooth pixels in the reference pixel sequence corresponding to the tooth pixel point may show a decreasing change pattern, which often indicates that the first i The more tooth pixels there are, the more likely they are to be plaque pixels.

[0114] In the third step, the plaque possible index corresponding to each tooth pixel is determined based on the plaque possible width, initial plaque suspected factor and reference suspected factor corresponding to each tooth pixel.

[0115] For example, the formula for determining the plaque indicator corresponding to a tooth pixel can be:

[0116] ;

[0117] in, It is i The plaque indicator corresponding to each tooth pixel. i is the serial number of the tooth pixel. It is i The possible width of plaque corresponding to each tooth pixel. It is i The initial plaque suspicion factor corresponding to each tooth pixel. It is i The plaque reference suspected factor corresponding to each tooth pixel.

[0118] It should be noted that when The larger the i The more plaque pixels there are in the preset horizontal window corresponding to the first tooth pixel, the more likely it is that there are more dental plaque pixels. i The more likely the pixel point of the tooth is to be a dental plaque pixel, and the i The greater the plaque level around each tooth pixel, the greater the The larger the iThe more likely the plaque width corresponding to the tooth pixel point is to be higher than the plaque width corresponding to the tooth pixel point in the corresponding reference pixel segment, the more likely the plaque width corresponding to the tooth pixel point is to be higher. i The more tooth pixels there are, the more likely they are to be dental plaque pixels. The larger the i The plaque widths corresponding to the adjacent tooth pixels in the reference pixel sequence corresponding to the tooth pixel point may show a decreasing change pattern, which often indicates that the first i The more tooth pixels there are, the more likely they are to be dental plaque pixels. The larger the i The more likely the first tooth pixel is to be a dental plaque pixel, the more likely the i The more tooth pixels there are, the more likely they are to show more plaque features.

[0119] Step S7: Based on the plaque possible indicators, all tooth surface images are fused, and dental plaque marker detection is performed based on the obtained fused image.

[0120] As an example, this step may include the following steps:

[0121] In the first step, the tooth pixels in all tooth surface images are fused according to the plaque characteristic expression values ​​and plaque possible indicators corresponding to the tooth pixels at the same position in all tooth surface images to obtain a fused image composed of all the fused tooth pixels.

[0122] For example, the formula for determining the grayscale value corresponding to the tooth pixel in the fused image can be:

[0123] ;

[0124] in, The fused image m The grayscale value corresponding to the tooth pixel. m The tooth pixel point can be the first m The pixel point is obtained by fusing the tooth pixels. m It is the serial number of the tooth pixel in the fused image or tooth surface image. M is the number of tooth surface images. f is the serial number of the tooth surface image. It is f In the tooth surface image m The plaque characteristic value corresponding to each tooth pixel. It is f In the tooth surface image m The plaque indicator corresponding to each tooth pixel. It is the first of all tooth surface imagesm The cumulative value of plaque possible indicators corresponding to the tooth pixel points. m The tooth pixel point and the tooth surface image m The tooth pixel can be a pixel at the same position in different images. m The tooth pixel is the pixel in the second row and fourth column of the fused image, so the tooth surface image has the m The tooth pixel point may also be the pixel point in the second row and fourth column in the tooth surface image.

[0125] It should be noted that when The larger the f In the tooth surface image m The more likely a tooth pixel is to show more plaque features, the more likely it is that the first f In the tooth surface image m The more tooth pixels there are, the more meaningful they are. The smaller the time, the more likely it is that f In the tooth surface image m The more tooth pixels there are, the more likely they are to be dental plaque pixels. Can characterize the first m The grayscale value of the pixel point obtained by fusing the tooth pixels is smaller, and the more likely it is a dental plaque pixel point.

[0126] In the second step, the fused image is segmented using the maximum inter-class variance method to obtain two sub-regions, and the sub-region with the smaller grayscale value in the two sub-regions is marked as the dental plaque region.

[0127] refer to Figure 2 Based on the same inventive concept as the above method embodiment, the present invention provides an oral dental plaque mark detection system, which includes a processor and a memory. The processor is used to process instructions stored in the memory to implement the steps of an oral dental plaque mark detection method, which may specifically include:

[0128] An acquisition and identification module 201 is used to acquire tooth surface images of a patient to be tested at different preset wavelengths and identify a target tooth area from each tooth surface image;

[0129] The plaque characteristic expression value determination module 202 is configured to determine the plaque characteristic expression value corresponding to each tooth pixel point in each target tooth region according to the grayscale value corresponding to each tooth pixel point and its corresponding preset wavelength;

[0130] The plaque possible width determination module 203 is configured to determine the plaque possible width corresponding to each tooth pixel point according to the plaque feature values corresponding to all tooth pixel points in the preset horizontal window corresponding to each tooth pixel point.

[0131] The ray and line segment construction module 204 is configured to take the tooth pixel point as an endpoint, extend a target ray in a vertical upward direction, and construct a tooth crown intersection line segment corresponding to the tooth pixel point according to the intersection of the target ray and the tooth crown region.

[0132] The pixel segment segmentation module 205 is configured to cut a reference pixel segment from the target ray corresponding to each tooth pixel point according to the tooth crown intersection line segment corresponding to each tooth pixel point.

[0133] The plaque possible index determination module 206 is configured to determine the plaque possible index corresponding to each tooth pixel point according to the difference between the plaque possible widths corresponding to the tooth pixel point in the reference pixel segment corresponding to each tooth pixel point.

[0134] The fusion and detection module 207 is configured to fuse all tooth surface images based on the plaque possible index, and detect the dental plaque markers based on the obtained fused image.

[0135] Figure 3 FIG. 1 is a structural schematic diagram of a computer device provided by an embodiment of the present application. As shown in the figure, the computer device 300 includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any of the above-described oral dental plaque marker detection methods. Figure 3

[0136] Based on the same inventive concept as the above method embodiments, the present application provides a server including a memory and a processor. The memory is configured to store executable program code, and the processor is configured to call and run the executable program code from the memory, so that the device executes any of the above-described oral dental plaque marker detection methods.

[0137] Based on the same inventive concept as the above method embodiments, the present application provides a computer program product including computer program code. When the computer program code runs on a computer, the computer executes any of the above-described oral dental plaque marker detection methods.

[0138] ​Based on the same inventive concept as the above-mentioned method embodiment, the present invention provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes any one of the above-mentioned oral plaque marking detection methods.

[0139] In summary, the present invention comprehensively analyzes the characteristic expressions of dental plaque in tooth surface images at different preset wavelengths, and specifically quantifies multiple indicators related to the characteristics of plaque expression, such as plaque characteristic expression value, plaque possible width and plaque possible index, thereby realizing the fusion of all tooth surface images, and realizing dental plaque marking detection based on the obtained fused image, thereby improving the accuracy of dental plaque marking detection of lower teeth, thereby improving the accuracy of dental plaque marking detection.

[0140] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A method for detecting oral plaque markers, characterized in that: The following steps are involved: Acquire tooth surface images of a patient to be tested at different preset wavelengths, and identify a target tooth region from each tooth surface image, wherein the tooth surface image is a hyperspectral image, a target tooth region represents a lower tooth, and the target tooth region includes a crown region; Determine the plaque characteristic value corresponding to each tooth pixel point in each target tooth area according to the grayscale value corresponding to each tooth pixel point and its corresponding preset wavelength; Determine the possible width of the plaque corresponding to each tooth pixel point based on the plaque characteristic expression values ​​corresponding to all tooth pixels within a preset horizontal window corresponding to each tooth pixel point; With the tooth pixel point as the endpoint, a target ray is drawn with the extension direction in the vertical upward direction, and the intersection of the target ray and the crown area forms the crown intersection line segment corresponding to the tooth pixel point; According to the tooth crown intersection line segment corresponding to each tooth pixel point, a reference pixel segment is cut out from the target ray corresponding to each tooth pixel point; Determining a plaque possible index corresponding to each tooth pixel point based on a difference between possible plaque widths corresponding to the tooth pixel point in a reference pixel segment corresponding to each tooth pixel point; Based on plaque possible indicators, all tooth surface images are fused and plaque marker detection is performed based on the resulting fused image; Determining the possible plaque index corresponding to each tooth pixel point according to the difference between the possible plaque widths corresponding to the tooth pixel point in the reference pixel segment corresponding to each tooth pixel point includes: Normalize the average of the difference between the plaque possible width corresponding to each tooth pixel and the plaque possible width corresponding to all tooth pixels in its corresponding reference pixel segment to obtain the initial plaque suspicion factor corresponding to each tooth pixel; Determine a plaque reference suspected factor corresponding to each tooth pixel point according to a difference between possible plaque widths corresponding to adjacent tooth pixels in a reference pixel segment corresponding to each tooth pixel point; The plaque possible index corresponding to each tooth pixel is determined based on the plaque possible width, initial plaque suspected factor and reference suspected factor corresponding to each tooth pixel.

2. The method for detecting oral plaque markers according to claim 1, wherein: The step of determining the plaque characteristic value corresponding to each tooth pixel point in each target tooth region according to the grayscale value corresponding to each tooth pixel point and its corresponding preset wavelength includes: Determine any tooth pixel as a marked pixel, and if the preset wavelength to which the marked pixel belongs is within a first preset wavelength range, determine the grayscale value corresponding to the marked pixel as the plaque characteristic expression value corresponding to the marked pixel; If the preset wavelength to which the marked pixel belongs is within the second preset wavelength range, the difference between the maximum value of the preset image grayscale range and the grayscale value corresponding to the marked pixel is determined as the plaque characteristic expression value corresponding to the marked pixel.

3. The method for detecting oral plaque markers according to claim 1, wherein: The determining of the possible plaque width corresponding to each tooth pixel point according to the plaque characteristic expression values ​​corresponding to all tooth pixels within a preset horizontal window corresponding to each tooth pixel point includes: The average of the plaque characteristic expression values ​​corresponding to all tooth pixels in all target tooth areas in each tooth surface image is determined as the overall characteristic expression value corresponding to each tooth surface image; The possible width of the plaque corresponding to each tooth pixel is determined based on the difference between the plaque feature expression values ​​corresponding to all tooth pixels within the preset horizontal window corresponding to each tooth pixel and the overall feature expression value corresponding to the tooth surface image to which they belong.

4. The method for detecting oral plaque markers according to claim 3, wherein: The determining of the possible plaque width corresponding to each tooth pixel point based on the difference between the plaque characteristic expression values ​​corresponding to all tooth pixels within a preset horizontal window corresponding to each tooth pixel point and the overall characteristic expression value corresponding to the tooth surface image to which they belong comprises: Determine any tooth pixel as a marked pixel, and determine the overall feature expression value corresponding to the tooth surface image to which the marked pixel belongs as the marked overall expression value; The average of the differences between the overall performance value of the mark and the plaque characteristic performance values ​​corresponding to each tooth pixel point in the preset horizontal window corresponding to the marked pixel point is normalized to obtain the possible width of the plaque corresponding to the marked pixel point.

5. The method for detecting oral plaque markers according to claim 1, wherein: The method of cutting out a reference pixel segment from a target ray corresponding to each tooth pixel point according to the tooth crown intersection segment corresponding to each tooth pixel point includes: Determine any tooth pixel as a marker pixel, and select the tooth pixel farthest from the marker pixel on the tooth crown intersection line segment corresponding to the marker pixel as the reference pixel corresponding to the marker pixel; The marked pixel point and its corresponding reference pixel point are connected to obtain a reference pixel segment corresponding to the marked pixel point.

6. The method for detecting oral plaque markers according to claim 1, wherein: The method of determining a plaque reference suspected factor corresponding to each tooth pixel point according to a difference between possible plaque widths corresponding to adjacent tooth pixel points in a reference pixel segment corresponding to each tooth pixel point includes: Sort the pixels in the reference pixel segment corresponding to each tooth pixel point in order from bottom to top to obtain a reference pixel sequence corresponding to each tooth pixel point; The mean of the difference between the possible plaque widths corresponding to all adjacent tooth pixels in the reference pixel sequence corresponding to each tooth pixel is normalized to obtain the plaque reference suspected factor corresponding to each tooth pixel.

7. The method for detecting oral plaque markers according to claim 1, wherein: The method of fusing all tooth surface images based on plaque possible indicators includes: According to the plaque characteristic expression values ​​and plaque possible indicators corresponding to the tooth pixels at the same position in all tooth surface images, the tooth pixels in all tooth surface images are fused to obtain a fused image composed of all the fused tooth pixels.

8. The method for detecting oral plaque markers according to claim 1, wherein: The detecting of dental plaque marks based on the obtained fused image includes: The obtained fused image is segmented by the maximum inter-class variance method to obtain two sub-regions, and the sub-region with the smaller gray value in the two sub-regions is marked as the dental plaque region.

9. An oral plaque marker detection system, characterized in that: The method comprises a processor and a memory, wherein the processor is used to process instructions stored in the memory to implement an oral plaque marking detection method according to any one of claims 1 to 8.

Images