CBCT image recognition method and system based on tooth calcification

Abstract

The invention discloses a CBCT image recognition method and system based on tooth calcification, and relates to the technical field of image recognition. Comprising the following steps: acquiring an oral cavity CBCT image; determining a tooth area image according to the oral cavity CBCT image; preprocessing the tooth area image; performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image; extracting edge calcification features of the enhanced tooth region image by using a canny operator; and identifying the extracted edge tooth calcification features through a U-net tooth calcification identification model. According to the method, the Canny operator is adopted to perform feature extraction, and the Canny is improved in a gradient magnitude calculation method, so that the Canny feature detection can be suitable for more scenes.

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and more particularly to a CBCT image recognition method and system based on tooth calcification. Background Art

[0002] With the continuous development and advancement of medical technology, Cone Beam Computer Tomography (CBCT) has been widely used in the diagnosis and treatment of oral diseases. CBCT images can show oral and maxillofacial structures such as teeth, tooth roots, jaw bone mass, jaw structure, alveolar bone, implants, apical resorption, cysts and tumors. It can be used for diagnosis and treatment planning in a variety of oral fields, such as bone mass assessment before implant surgery, position and morphology assessment before wisdom tooth extraction, alveolar cleft detection, diagnosis of temporomandibular joint diseases, etc.

[0003] The existing computed tomography (CT) images have a long scanning time, strong radiation, low image spatial resolution, and poor precision recognition effect. Compared with the existing CT scans, CBCT images are more suitable for applications in the oral field. It can provide dentists with more comprehensive oral structure information, so as to formulate treatment plans more accurately and improve treatment effects.

[0004] Therefore, it is an urgent problem for those skilled in the art to propose a CBCT image recognition method and system based on tooth calcification to solve the difficulties existing in the prior art. Summary of the invention

[0005] In view of this, the present invention provides a CBCT image recognition method and system based on tooth calcification, which are used to solve the technical problems existing in the prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A CBCT image recognition method based on tooth calcification comprises the following steps:

[0008] Acquire oral CBCT images;

[0009] Determine a tooth region image based on an oral CBCT image;

[0010] Preprocessing the tooth area image;

[0011] Performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image;

[0012] The canny operator is used to extract the edge calcification features of the enhanced tooth area image;

[0013] The extracted marginal tooth calcification features were identified using the U-net tooth calcification recognition model.

[0014] The above method may optionally utilize a CBCT scanner to acquire oral CBCT images.

[0015] In the above method, optionally, the specific content of the tooth region image is determined based on the oral CBCT image as follows:

[0016] Projecting the oral CBCT image along the coronal plane of the teeth in the oral cavity to determine a target projection image corresponding to the oral CBCT image;

[0017] According to the target projection image, a maximum density projection method is used to detect the tooth region in the target projection image to obtain a target layer of the tooth region;

[0018] A tooth region image is determined according to a target layer of the tooth region.

[0019] In the above method, optionally, the specific content of preprocessing the tooth area image is:

[0020] Scanning the acquired CBCT image to obtain the HU value of the CBCT image;

[0021] Limiting the HU value of the CBCT image to obtain a CBCT image after the limit processing;

[0022] The CBCT images after restriction processing were processed using window technology and normalization technology.

[0023] In the above method, optionally, the specific contents of image enhancement on the preprocessed tooth region image are as follows:

[0024] Histogram equalization, adaptive histogram equalization and contrast enhancement methods are used for image enhancement.

[0025] The above method may optionally use the canny operator to extract the edge calcification features of the enhanced tooth region image as follows:

[0026] The enhanced tooth region image is smoothed using a Gaussian filter;

[0027] Calculate the gradient magnitude;

[0028] Perform non-maximum suppression on the gradient amplitude;

[0029] Detect and connect edges using a dual threshold algorithm.

[0030] In the above method, optionally, the specific content of calculating the gradient amplitude is:

[0031]

[0032] Among them, h(x,y) is the gradient of the pixel point (x,y), h1(x,y), h2(x,y), h3(x,y), h4(x,y) are the gradients of the four directions of the pixel point, and the gradient directions are

[0033] In the above method, optionally, the specific content of performing non-maximum suppression on the gradient amplitude is:

[0034] Use the Canny operator 3×3 template to find the maximum value of the gradient amplitude;

[0035] Determine whether the search result is greater than the gradient amplitude in the adjacent gradient direction;

[0036] If the judgment result is yes, the search result is a maximum value;

[0037] If the judgment result is no, the search result is set to 0, and the search continues until the maximum value of the gradient amplitude is found.

[0038] In the above method, optionally, the U-net tooth calcification recognition model includes:

[0039] The U-net tooth calcification recognition model adopts an encoder-decoder structure;

[0040] The encoder applies multi-scale feature fusion. Each multi-scale feature fusion module contains multiple convolution paths, and each convolution path consists of one or more convolution operations.

[0041] Multi-scale skip connections are applied to the decoder modules, and each decoder receives feature images from the previous layer as well as the encoder and decoder of the same layer and the next layer.

[0042] A CBCT image recognition system based on tooth calcification, applying any one of the above-mentioned CBCT image recognition methods based on tooth calcification, comprising an image acquisition module, a region determination module, a preprocessing module, an image enhancement module, a feature extraction module, and a feature recognition module connected in sequence;

[0043] Image acquisition module: acquire oral CBCT images;

[0044] Determine region module: determine the tooth region image based on the oral CBCT image;

[0045] Preprocessing module: preprocess the tooth area image;

[0046] Image enhancement module: performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image;

[0047] Feature extraction module: Use the canny operator to extract the edge calcification features of the enhanced tooth area image;

[0048] Feature recognition module: The extracted marginal tooth calcification features are recognized through the U-net tooth calcification recognition model.

[0049] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a CBCT image recognition method and system based on tooth calcification, which has the following beneficial effects:

[0050] 1) The present invention eliminates subjective error factors related to the environment, instruments and equipment, and saves time and effort;

[0051] 2) The Canny operator is used for feature extraction and the Canny gradient amplitude calculation method is improved, so that the Canny feature detection can be applied to more scenarios;

[0052] 3) Multi-scale feature fusion was added to the encoder to provide richer feature information; multi-scale jump connections were added to the decoder so that each decoder could obtain feature information of all scales, and the U-net tooth calcification recognition model was used to improve the efficiency of tooth calcification recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0054] Figure 1 A flowchart of a CBCT image recognition method based on tooth calcification provided by the present invention;

[0055] Figure 2 This is a structural diagram of a CBCT image recognition system based on tooth calcification provided by the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] See also Figure 1As shown, the present invention discloses a CBCT image recognition method based on tooth calcification, comprising the following steps:

[0058] Acquire oral CBCT images;

[0059] Determine a tooth region image based on an oral CBCT image;

[0060] Preprocessing the tooth area image;

[0061] Performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image;

[0062] The canny operator is used to extract the edge calcification features of the enhanced tooth area image;

[0063] The extracted marginal tooth calcification features were identified using the U-net tooth calcification recognition model.

[0064] Furthermore, oral CBCT images are acquired using a CBCT scanner.

[0065] Specifically, the image resolution of the CBCT scanner is 0.2-1.0 mm, and the CBCT volume size is 672×688×688.

[0066] Before using a CBCT scanner to scan the teeth in the mouth, the gums and soft tissues in the mouth need to be separated first to make the image scanned by the CBCT scanner clearer and more accurate. Therefore, an opener is needed to separate the gums and soft tissues to meet the needs of the actual scanning operation. Since the patient is in an open-mouthed state when using an opener, the CBCT scanner obtains oral CBCT image information in the open-mouthed state. It is also necessary to collect a set of oral CBCT image information in the upper and lower jaw occlusal state. By integrating and correcting the two sets of image information, the obtained image information is clearer and more complete, and the positional relationship of each tissue in the obtained three-dimensional oral model is also more accurate.

[0067] During the acquisition process, images can be acquired using cone-beam CBCT or spiral CT. Both methods can obtain oral CBCT images.

[0068] Furthermore, the specific content of the tooth area image is determined based on the oral CBCT image:

[0069] Projecting the oral CBCT image along the coronal plane of the teeth in the oral cavity to determine a target projection image corresponding to the oral CBCT image;

[0070] According to the target projection image, a maximum density projection method is used to detect the tooth region in the target projection image to obtain a target layer of the tooth region;

[0071] A tooth region image is determined according to a target layer of the tooth region.

[0072] Specifically, the maximum density projection method is to perform maximum density projection on the CBCT image to obtain the maximum density projection points of the CBCT image along different directions. Further maximum response detection is performed on these projection points to determine the position information of the tooth area in various directions, thereby accurately identifying the tooth area.

[0073] Furthermore, the specific contents of preprocessing the tooth area image are as follows:

[0074] Scanning the acquired CBCT image to obtain the HU value of the CBCT image;

[0075] Limiting the HU value of the CBCT image to obtain a CBCT image after the limit processing;

[0076] The CBCT images after restriction processing were processed using window technology and normalization technology.

[0077] Specifically, window technology is a key technology in the field of medical imaging. It is a display technology used to observe normal tissues or lesions of different densities in CT examinations, including window width and window level. Since various tissue structures have different CT values, when you want to display the details of a certain tissue structure, you should choose a window width and window level suitable for observing the tissue to obtain the best display. The window width is the CT value range displayed on the CT image, and the tissues within this CT value range are displayed in different simulated grayscales. The window level is the center position of the window. For the same window width, due to different window levels, the CT values ​​of the CT value range included are also different.

[0078] Standardization technology is to centralize the image by removing the mean. According to convex optimization theory and knowledge related to data probability distribution, data centralization conforms to the law of data distribution and is more likely to achieve generalization effects after training.

[0079] Furthermore, the specific contents of image enhancement for the preprocessed tooth region image are as follows:

[0080] Histogram equalization, adaptive histogram equalization and contrast enhancement methods are used for image enhancement.

[0081] Specifically, these methods aim to enhance the contrast and details of the image by adjusting the grayscale value of the image to make the tooth area more clearly visible.

[0082] Histogram equalization: This method adjusts the grayscale value of the image and adjusts the histogram by stretching the contrast, thereby increasing the grayscale difference between the foreground and background to achieve the purpose of enhancing the contrast.

[0083] Adaptive histogram equalization: By limiting the maximum slope of the cumulative distribution function (CDF), the limitations of the histogram equalization algorithm are overcome, the random noise introduced in the histogram equalization process is eliminated, and the noise can be limited to a certain extent and the amplification of boundary problems can be improved.

[0084] Contrast Enhancement Method: Further optimizes the effect of image enhancement to improve visibility and clarity of the tooth area.

[0085] Furthermore, the specific content of using the canny operator to extract the edge calcification features of the enhanced tooth area image is:

[0086] The enhanced tooth region image is smoothed using a Gaussian filter;

[0087] Calculate the gradient magnitude;

[0088] Perform non-maximum suppression on the gradient amplitude;

[0089] Detect and connect edges using a dual threshold algorithm.

[0090] Specifically, a Gaussian filter is used to perform smoothing filtering and eliminate outliers.

[0091] Furthermore, the specific content of calculating the gradient amplitude is:

[0092]

[0093] Among them, h(x,y) is the gradient of the pixel point (x,y), h1(x,y), h2(x,y), h3(x,y), h4(x,y) are the gradients of the four directions of the pixel point, and the gradient directions are

[0094] Furthermore, the specific content of non-maximum suppression of gradient amplitude is:

[0095] Use the Canny operator 3×3 template to find the maximum value of the gradient amplitude;

[0096] Determine whether the search result is greater than the gradient amplitude in the adjacent gradient direction;

[0097] If the judgment result is yes, the search result is a maximum value;

[0098] If the judgment result is no, the search result is set to 0, and the search continues until the maximum value of the gradient amplitude is found.

[0099] Furthermore, the U-net tooth calcification recognition model includes:

[0100] The U-net tooth calcification recognition model adopts an encoder-decoder structure;

[0101] The encoder applies multi-scale feature fusion. Each multi-scale feature fusion module contains multiple convolution paths, and each convolution path consists of one or more convolution operations.

[0102] Specifically, the shape of the convolution kernel can be 1*1, 3*3, 5*5 and 7*7.

[0103] The formula for multi-scale feature fusion is expressed as follows:

[0104]

[0105] in, and are the input and output of the i-th encoder, respectively. and are the convolution operations of path a and path b of the i-th encoder, respectively, and C out It is a 3*3 convolution operation.

[0106] Multi-scale skip connections are applied to the decoder modules, and each decoder receives feature images from the previous layer as well as the encoder and decoder of the same layer and the next layer.

[0107] Specifically, the expression of multi-scale skip connection is:

[0108]

[0109] Among them, De i is the i-th decoder, En i is the i-th encoder, is a feature fusion operation, which is composed of a 3*3 convolution operation, batch normalization and ReLU activation function in series. [·] is a feature concatenation operation. C(·) They are convolution operation, downsampling and upsampling operation respectively.

[0110] and Figure 1 Corresponding to the method described above, the present invention also provides a CBCT image recognition system based on tooth calcification for Figure 1 The specific implementation of the method in the above example is shown in the schematic diagram of its structure. Figure 2 As shown, it includes an image acquisition module, a region determination module, a preprocessing module, an image enhancement module, a feature extraction module, and a feature recognition module connected in sequence;

[0111] Image acquisition module: acquire oral CBCT images;

[0112] Determine region module: determine the tooth region image based on the oral CBCT image;

[0113] Preprocessing module: preprocess the tooth area image;

[0114] Image enhancement module: performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image;

[0115] Feature extraction module: Use the canny operator to extract the edge calcification features of the enhanced tooth area image;

[0116] Feature recognition module: The extracted marginal tooth calcification features are recognized through the U-net tooth calcification recognition model.

[0117] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0118] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A CBCT image recognition method based on tooth calcification, characterized in that: The following steps are involved: Acquire oral CBCT images; Determine a tooth region image based on an oral CBCT image; Preprocessing the tooth area image; Performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image; The canny operator is used to extract the edge calcification features of the enhanced tooth area image; The extracted marginal tooth calcification features were identified using the U-net tooth calcification recognition model.

2. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: The oral CBCT images were acquired using a CBCT scanner.

3. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: The specific contents of the tooth area image determined based on the oral CBCT image are: Projecting the oral CBCT image along the coronal plane of the teeth in the oral cavity to determine a target projection image corresponding to the oral CBCT image; According to the target projection image, a maximum density projection method is used to detect the tooth region in the target projection image to obtain a target layer of the tooth region; A tooth region image is determined according to a target layer of the tooth region.

4. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: The specific contents of preprocessing the tooth area image are as follows: Scanning the acquired CBCT image to obtain the HU value of the CBCT image; Limiting the HU value of the CBCT image to obtain a CBCT image after the limit processing; The CBCT images after restriction processing were processed using window technology and normalization technology.

5. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: The specific contents of image enhancement for the preprocessed tooth area image are as follows: Histogram equalization, adaptive histogram equalization and contrast enhancement methods are used for image enhancement.

6. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: The specific content of using the canny operator to extract the edge calcification features of the enhanced tooth area image is: The enhanced tooth region image is smoothed using a Gaussian filter; Calculate the gradient magnitude; Perform non-maximum suppression on the gradient amplitude; Detect and connect edges using a dual threshold algorithm.

7. The CBCT image recognition method based on tooth calcification according to claim 6, characterized in that: The specific content of calculating the gradient amplitude is: Among them, h(x,y) is the gradient of the pixel point (x,y), h1(x,y), h2(x,y), h3(x,y), h4(x,y) are the gradients of the four directions of the pixel point, and the gradient directions are 8. The CBCT image recognition method based on tooth calcification according to claim 6, characterized in that: The specific content of non-maximum suppression of gradient amplitude is: Use the Canny operator 3×3 template to find the maximum value of the gradient amplitude; Determine whether the search result is greater than the gradient amplitude in the adjacent gradient direction; If the judgment result is yes, the search result is a maximum value; If the judgment result is no, the search result is set to 0, and the search continues until the maximum value of the gradient amplitude is found.

9. The CBCT image recognition method based on tooth calcification according to claim 1, characterized in that: U-net tooth calcification recognition model includes: The U-net tooth calcification recognition model adopts an encoder-decoder structure; The encoder applies multi-scale feature fusion. Each multi-scale feature fusion module contains multiple convolution paths, and each convolution path consists of one or more convolution operations. Multi-scale skip connections are applied to the decoder modules, and each decoder receives feature images from the previous layer as well as the encoder and decoder of the same layer and the next layer.

10. A CBCT image recognition system based on tooth calcification, characterized in that A CBCT image recognition method based on tooth calcification according to any one of claims 1 to 9 is applied, comprising an image acquisition module, a region determination module, a preprocessing module, an image enhancement module, a feature extraction module, and a feature recognition module connected in sequence; Image acquisition module: acquire oral CBCT images; Determine region module: determine the tooth region image based on the oral CBCT image; Preprocessing module: preprocess the tooth area image; Image enhancement module: performing image enhancement on the preprocessed tooth region image to obtain an enhanced tooth region image; Feature extraction module: Use the canny operator to extract the edge calcification features of the enhanced tooth area image; Feature recognition module: The extracted marginal tooth calcification features are recognized through the U-net tooth calcification recognition model.

Images