Bronchoscope positioning method and device

Abstract

The invention relates to a bronchoscope positioning method and device. The method comprises the following steps: acquiring a medical image acquired by a bronchoscope; performing normalization processing on the medical image to obtain a normalized medical image; feature points on the normalized medical image are screened out; extracting feature points on the normalized medical image to form a feature image; inputting the feature image into a neural network for training to obtain a positioning recognition model; and identifying the type and the position of the target medical image by using the positioning identification model. According to the method, the medical examination image is subjected to normalization processing and feature point extraction, so that the quality and features of the medical image can be enhanced, the target tissue can be recognized more accurately, the accuracy of abnormal tissue positioning is improved, and the risk of missed diagnosis or misdiagnosis is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of bronchoscope positioning, and in particular to a bronchoscope positioning method and device. Background Art

[0002] Bronchoscopy is an important clinical diagnostic and treatment method, mainly used to observe, sample and treat the lesions in the airway. With the continuous development of medical imaging technology, the clarity and resolution of bronchoscopic images have been significantly improved, allowing doctors to observe abnormal tissues in the airway more accurately. However, traditional bronchoscopy usually relies on the doctor's experience to locate and sample abnormal tissues, which may make it difficult to accurately find the lesions, resulting in missed diagnosis or misdiagnosis. Summary of the invention

[0003] In order to solve the above problems, an object of the embodiments of the present invention is to provide a bronchoscope positioning method and device.

[0004] A bronchoscope positioning method, comprising:

[0005] Step 1: Obtain medical images collected by bronchoscope;

[0006] Step 2: performing normalization processing on the medical image to obtain a normalized medical image;

[0007] Step 3: Filter out the feature points on the normalized medical image;

[0008] Step 4: Extract feature points on the normalized medical image to form a feature image;

[0009] Step 5: Input the feature image into the neural network for training to obtain a positioning recognition model;

[0010] Step 6: Use the localization recognition model to identify the type and location of the target medical image.

[0011] Preferably, the step 2: normalizing the medical image to obtain a normalized medical image includes:

[0012] Step 2.1: grayscale the medical image to obtain a grayscale medical image; wherein the grayscale processing formula is:

[0013] Gray=RED*0.299+GREEN*0.587+BLUE*0.114

[0014] Among them, Gray represents the grayscale medical image, RED represents the value of the medical image in the R channel, GREEN represents the value of the medical image in the G channel, and BLUE represents the value of the medical image in the B channel;

[0015] Step 2.2: performing standardization processing on the grayscale medical image to obtain a standardized medical image;

[0016] Step 2.3: Scale the standardized medical image to obtain a normalized medical image.

[0017] Preferably, in step 2.3, the scaling calculation formula is:

[0018]

[0019] Among them, G represents the pixel value on the normalized medical image, g represents the standardized medical image, w1 represents the first scaling factor, w2 represents the second scaling factor, input represents the input pixel value of the grayscale medical image, mean represents the mean of the grayscale medical image, and std represents the standard deviation of the grayscale medical image.

[0020] Preferably, the step 3: screening out feature points on the normalized medical image includes:

[0021] Step 3.1: Calculate the gradient value of each pixel in the normalized medical image in the x direction and the y direction;

[0022] Step 3.2: Use the gradient value to construct a feature point extraction matrix;

[0023] Step 3.3: constructing a feature point judgment model based on the feature point extraction matrix;

[0024] Step 3.4: Calculate the decision value of each pixel using the feature point judgment model;

[0025] Step 3.5: Pixels whose judgment values ​​are greater than the preset value are taken as feature points.

[0026] Preferably, the step 3.1: calculating the gradient value of each pixel in the normalized medical image in the x direction and the y direction comprises:

[0027] The gradient value of each pixel in the x and y directions is solved by using the difference of adjacent pixel points; the gradient value calculation formula is:

[0028]

[0029] Among them, I x Indicates the gradient value of pixel I(x,y) in the x direction, I y It represents the gradient value of pixel point I(x,y) in the y direction, and I(x,y) represents the pixel value at the (x,y) position.

[0030] Preferably, the step 3.2: constructing a feature point extraction matrix using gradient values ​​includes:

[0031] Use the Gaussian function to convolve with the gradient value to construct a feature point extraction matrix; where the feature point extraction matrix is:

[0032]

[0033] Among them, M represents the feature point extraction matrix, represents the convolution operation, G represents the Gaussian function, σ represents the Gaussian scale parameter.

[0034] Preferably, in step 3.3, the feature point criterion model is:

[0035] R=detM-τ(traceM) 2

[0036] Among them, R represents the decision value, detM represents the determinant of the feature point extraction matrix, traceM represents the direct product of the feature point extraction matrix, and τ represents the preset coefficient.

[0037] The present invention also provides a bronchoscope positioning device, comprising:

[0038] An image acquisition module, used for acquiring medical images collected by a bronchoscope;

[0039] A normalization module, used for performing normalization processing on the medical image to obtain a normalized medical image;

[0040] A feature point screening module is used to screen out feature points on the normalized medical image;

[0041] A feature image extraction module is used to extract feature points on the normalized medical image to form a feature image;

[0042] A training module is used to input the feature image into the neural network for training to obtain a positioning recognition model;

[0043] The positioning recognition module is used to use the positioning recognition model to identify the type and location of the target medical image.

[0044] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and wherein the computer program, when executed by the processor, implements the steps in the above-mentioned bronchoscope positioning method.

[0045] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps in the above-mentioned bronchoscope positioning method when executed by a processor.

[0046] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0047] The present invention relates to a bronchoscope positioning method. Compared with the prior art, the present invention can enhance the quality and features of medical images by normalizing and extracting feature points of medical examination images, making the identification of target tissue more accurate, thereby improving the accuracy of abnormal tissue positioning and reducing the risk of missed diagnosis or misdiagnosis.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 A flow chart of a bronchoscope positioning method provided by the present invention;

[0051] Figure 2 This is a schematic diagram of a bronchoscope positioning device provided by the present invention. DETAILED DESCRIPTION

[0052] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0053] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0054] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] See also Figure 1 , a bronchoscope positioning method, comprising:

[0056] Step 1: Obtain medical images collected by bronchoscope;

[0057] Step 2: performing normalization processing on the medical image to obtain a normalized medical image;

[0058] Furthermore, the step 2 comprises:

[0059] Step 2.1: grayscale the medical image to obtain a grayscale medical image; wherein the grayscale processing formula is:

[0060] Gray=RED*0.299+GREEN*0.587+BLUE*0.114

[0061] Among them, Gray represents the grayscale medical image, RED represents the value of the medical image in the R channel, GREEN represents the value of the medical image in the G channel, and BLUE represents the value of the medical image in the B channel;

[0062] Step 2.2: performing standardization processing on the grayscale medical image to obtain a standardized medical image;

[0063] Step 2.3: scaling the standardized medical image to obtain a normalized medical image;

[0064] In step 2.3, the present invention may adopt the scaling calculation formula:

[0065]

[0066] The standardized medical image is scaled to obtain a normalized medical image; wherein G represents a pixel value on the normalized medical image, g represents a standardized medical image, w1 represents a first scaling factor, w2 represents a second scaling factor, input represents an input pixel value of the grayscale medical image, mean represents a mean of the grayscale medical image, and std represents a standard deviation of the grayscale medical image.

[0067] The present invention can make the network pay more attention to the pixels in the dark area during training by performing scaling processing on the image, because the dark area usually represents a greater possibility of the existence of abnormal tissue.

[0068] Step 3: Filter out the feature points on the normalized medical image;

[0069] Furthermore, step 3 includes:

[0070] Step 3.1: Calculate the gradient value of each pixel in the normalized medical image in the x direction and the y direction;

[0071] Among them, step 3.1 includes:

[0072] The gradient value of each pixel in the x and y directions is solved by using the difference of adjacent pixel points; the gradient value calculation formula is:

[0073]

[0074] Among them, I x Indicates the gradient value of pixel I(x,y) in the x direction, I y It represents the gradient value of pixel point I(x,y) in the y direction, and I(x,y) represents the pixel value at the (x,y) position.

[0075] Step 3.2: Use the gradient value to construct a feature point extraction matrix;

[0076] In step 3.2, the present invention may use a Gaussian function and a gradient value for convolution to construct a feature point extraction matrix; wherein the feature point extraction matrix is:

[0077]

[0078] Among them, M represents the feature point extraction matrix, represents the convolution operation, G represents the Gaussian function, σ represents the Gaussian scale parameter.

[0079] Step 3.3: constructing a feature point judgment model based on the feature point extraction matrix;

[0080] Step 3.4: Calculate the decision value of each pixel using the feature point judgment model;

[0081] In step 3.4, the feature point criterion model is:

[0082] R=detM-τ(traceM) 2

[0083] Among them, R represents the decision value, detM represents the determinant of the feature point extraction matrix, traceM represents the direct product of the feature point extraction matrix, and τ represents the preset coefficient.

[0084] Step 3.5: Pixels whose judgment values ​​are greater than the preset value are taken as feature points.

[0085] The gradient-based feature screening method of the present invention can not only suppress noise interference in the image to a certain extent, but also enable the neural network to focus on the feature information in the image, which helps the model learn more representative features and improves the generalization ability of the model.

[0086] Step 4: Extract feature points on the normalized medical image to form a feature image;

[0087] Step 5: Input the feature image into the neural network for training to obtain a positioning recognition model;

[0088] Step 6: Use the localization recognition model to identify the type and location of the target medical image.

[0089] The present invention can enhance the quality and features of medical images by normalizing and extracting feature points of medical examination images, making the identification of target tissue more accurate, thereby improving the accuracy of abnormal tissue positioning and reducing the risk of missed diagnosis or misdiagnosis.

[0090] See also Figure 2 The present invention also provides a bronchoscope positioning device, comprising:

[0091] An image acquisition module, used for acquiring medical images collected by a bronchoscope;

[0092] A normalization module, used for performing normalization processing on the medical image to obtain a normalized medical image;

[0093] A feature point screening module is used to screen out feature points on the normalized medical image;

[0094] A feature image extraction module is used to extract feature points on the normalized medical image to form a feature image;

[0095] A training module is used to input the feature image into the neural network for training to obtain a positioning recognition model;

[0096] The positioning recognition module is used to use the positioning recognition model to identify the type and location of the target medical image.

[0097] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory and the processor are connected via the bus, and is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned bronchoscope positioning method are implemented. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as the beneficial effects of the bronchoscope positioning method described in the above-mentioned technical solution, and are not elaborated herein.

[0098] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned bronchoscope positioning method are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the bronchoscope positioning method described in the above-mentioned technical solution, which will not be repeated here.

[0099] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A bronchoscope positioning method, characterized in that: include: Step 1: Obtain medical images collected by bronchoscope; Step 2: performing normalization processing on the medical image to obtain a normalized medical image; Step 3: Filter out the feature points on the normalized medical image; Step 4: Extract feature points on the normalized medical image to form a feature image; Step 5: Input the feature image into the neural network for training to obtain a positioning recognition model; Step 6: Use the localization recognition model to identify the type and location of the target medical image.

2. A bronchoscope positioning method according to claim 1, characterized in that: The step 2: performing normalization processing on the medical image to obtain a normalized medical image includes: Step 2.1: grayscale the medical image to obtain a grayscale medical image; wherein the grayscale processing formula is: Gray=RED*0.299+GREEN*0.587+BLUE*0.114 Among them, Gray represents the grayscale medical image, RED represents the value of the medical image in the R channel, GREEN represents the value of the medical image in the G channel, and BLUE represents the value of the medical image in the B channel; Step 2.2: performing standardization processing on the grayscale medical image to obtain a standardized medical image; Step 2.3: Scale the standardized medical image to obtain a normalized medical image.

3. A bronchoscope positioning method according to claim 2, characterized in that: In step 2.3, the scaling calculation formula is: Among them, G represents the pixel value on the normalized medical image, g represents the standardized medical image, w1 represents the first scaling factor, w2 represents the second scaling factor, input represents the input pixel value of the grayscale medical image, mean represents the mean of the grayscale medical image, and std represents the standard deviation of the grayscale medical image.

4. A bronchoscope positioning method according to claim 1, characterized in that: The step 3: screening out feature points on the normalized medical image includes: Step 3.1: Calculate the gradient value of each pixel in the normalized medical image in the x direction and the y direction; Step 3.2: Use the gradient value to construct a feature point extraction matrix; Step 3.3: constructing a feature point judgment model based on the feature point extraction matrix; Step 3.4: Calculate the decision value of each pixel using the feature point judgment model; Step 3.5: Pixels whose judgment values ​​are greater than the preset value are taken as feature points.

5. A bronchoscope positioning method according to claim 4, characterized in that: The step 3.1: calculating the gradient value of each pixel in the normalized medical image in the x direction and the y direction, comprises: The gradient value of each pixel in the x and y directions is solved by using the difference of adjacent pixel points; the gradient value calculation formula is: Among them, I x Indicates the gradient value of pixel I(x,y) in the x direction, I y It represents the gradient value of pixel point I(x,y) in the y direction, and I(x,y) represents the pixel value at the (x,y) position.

6. A bronchoscope positioning method according to claim 5, characterized in that: The step 3.2: constructing a feature point extraction matrix using gradient values, includes: Use the Gaussian function to convolve with the gradient value to construct a feature point extraction matrix; where the feature point extraction matrix is: Among them, M represents the feature point extraction matrix, represents the convolution operation, G represents the Gaussian function, σ represents the Gaussian scale parameter.

7. A bronchoscope positioning method according to claim 6, characterized in that: In step 3.3, the feature point criterion model is: R=detM-τ(traceM) 2 Among them, R represents the decision value, detM represents the determinant of the feature point extraction matrix, traceM represents the direct product of the feature point extraction matrix, and τ represents the preset coefficient.

8. A bronchoscope positioning device, characterized in that: include: An image acquisition module, used for acquiring medical images collected by a bronchoscope; A normalization module, used for performing normalization processing on the medical image to obtain a normalized medical image; A feature point screening module is used to screen out feature points on the normalized medical image; A feature image extraction module is used to extract feature points on the normalized medical image to form a feature image; A training module is used to input the feature image into the neural network for training to obtain a positioning recognition model; The positioning recognition module is used to use the positioning recognition model to identify the type and location of the target medical image.

9. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps of a bronchoscope positioning method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a bronchoscope positioning method as described in any one of claims 1-7 are implemented.

Images