Oral cavity scanning device

Abstract

The utility model provides an oral cavity scanning device which comprises a glass cover body, a cavity is formed in the glass cover body, the glass cover body extends in the length direction, a motor is arranged on the end face of one end of the glass cover body, an output shaft of the motor is connected with a rod body, and the rod body is arranged in the length direction of the glass cover body. And an image acquisition head, a blue laser light source and a white light LED are arranged on the surface of the rod body. A memory and / or a wireless transmission device are / is arranged in the main control device; the main control device is electrically connected with the motor, the image acquisition head, the blue laser light source and the white light LED; and a power supply is also arranged for supplying power. According to the utility model, images in the oral cavity can be rapidly acquired in various modes, and guidance is provided for subsequent acquisition and identification of oral cavity pathological tissue images.

Description

Technical Field

[0001] The utility model relates to the field of oral image screening, in particular to an oral scanning device used for oral image screening. Background Art

[0002] Oral cell cancer is a common malignant tumor. Early detection and treatment have an important impact on the patient's survival rate and quality of life. Image screening is a rapid and non-invasive detection method that can be used to assist doctors in diagnosing oral cancer. Chinese patent document CN 111563887 A records a method and device for intelligent analysis of oral images based on vision, but the scheme is mainly based on conventional visual images for analysis. It fails to reach microscopic tissue and cell images. "Research on Oral Cancer Image Recognition Based on Deep Learning" by Jiao Long records a scheme for oral cancer pathological image recognition based on convolutional neural networks, and introduces various different models, but these models are usually large and the recognition processing time is long. And the models are all trained based on high-quality image data. When used in actual image recognition, the actual application effect of these models is poor due to the large number of unsatisfactory images collected in clinical images. The technical problem further discovered is that oral epithelial cells are polygonal or rectangular and are about 20-30 microns in size. The cytoplasm contains many small granular substances, called organelles, including mitochondria, endoplasmic reticulum, Golgi bodies, etc. There are many glycoprotein molecules on the outside of the cell membrane, called adhesion molecules, which play a role in interacting with neighboring cells or tissues. In addition, some unstained substances can be seen, which may be cell nuclei or cytoplasmic fragments. These will also interfere with the normal operation of the model during image recognition, resulting in a decrease in recognition accuracy. Moreover, the scope of the oral cavity is large, and sampling of tissues at different locations may result in large differences in recognition results. Clinically, cases have been found in which the results obtained by sampling patients at different locations in the oral cavity are completely different. Due to the large scope of the oral cavity, more images need to be collected. How to quickly locate the abnormal position of the patient's oral tissue and improve the accuracy of image acquisition is difficult. The above technical difficulties have hindered the development of oral pathology recognition and increased the difficulty of early detection of oral cancer. It should be noted that the records in the background technology are not an admission of the prior art. The inventors raise new technical issues, which are also part of their innovative contributions. Utility Model Content

[0003] The technical problem to be solved by the utility model is to provide an oral cavity scanning device, which can assist in improving the efficiency and accuracy of oral cavity image screening.

[0004] In order to solve the above technical problems, the technical solution adopted by the utility model is: an oral scanning device, including a glass cover body, the glass cover body has a cavity, the glass cover body extends along the length direction, a motor is provided on the end face of one end of the glass cover body, the output shaft of the motor is connected to the rod body, the rod body is arranged along the length direction of the glass cover body, and an image acquisition head, a blue laser light source and a white light LED are provided on the surface of the rod body.

[0005] In the preferred solution, a main control device is also provided, and a memory and / or a wireless transmission device are provided in the main control device;

[0006] The main control device is electrically connected to the motor, the image acquisition head, the blue laser light source and the white light LED;

[0007] A power supply is also provided for supplying power.

[0008] In a preferred embodiment, a co-convergent laser light source is provided on the rod body along the length direction. The co-convergent laser light source is an array of laser light sources, which are columns along the length direction and rows along the circumferential direction. At least multiple lasers in the same row are focused on the same focal point.

[0009] In a preferred embodiment, the number of the co-convergent laser light sources is 2 to 8, and the co-convergent laser light sources are located on the outer wall of the rod body.

[0010] In a preferred solution, the co-convergent laser light sources are symmetrically arranged on both sides of the image acquisition head, the blue laser light source and the white light LED.

[0011] In a preferred solution, the image acquisition head, the blue laser light source and the white light LED are arranged in a row.

[0012] In a preferred solution, the image acquisition head is located in the middle of the rod body.

[0013] In a preferred solution, there are at least two blue laser light sources and at least two white light LEDs, and the blue laser light sources and the white light LEDs are symmetrically arranged near the two ends of the image acquisition head.

[0014] A heating device is also provided in the glass cover body, and the heating device is electrically connected to the main control device to heat the oral scanning device to human body temperature to prevent water vapor from being generated in the glass cover body to affect the field of vision.

[0015] In a preferred solution, the cross section of the glass cover is a triangle, and the edges of the triangle are chamfered.

[0016] In a preferred solution, an opening is provided on one end surface of the glass cover body, a base is provided at the position of the opening, and the motor is fixed on the base.

[0017] The utility model provides an oral scanning device that can quickly collect oral images in multiple modes, providing guidance for the subsequent collection and recognition of oral pathological tissue images. In addition, the multi-dimensional oral images are also used as weight parameters for subsequent oral pathological tissue image recognition, further improving the accuracy of the final result. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The utility model is further described below in conjunction with the accompanying drawings and embodiments:

[0019] Figure 1 It is a structural schematic diagram of the utility model.

[0020] Figure 2 Oral epithelial cell images collected for guidance of the present invention.

[0021] In the figure: motor 1, glass cover 2, image acquisition head 3, co-convergent laser light source 4, blue laser light source 5, white light LED 6, rod body 7. DETAILED DESCRIPTION

[0022] Embodiment 1:

[0023] like Figure 1 In the present invention, an oral scanning device includes a glass cover 2, a cavity is provided in the glass cover 2, the glass cover 2 extends in the length direction, a motor 1 is provided at one end face of the glass cover 2, an output shaft of the motor 1 is connected to a rod 7, the rod 7 is arranged along the length direction of the glass cover 2, and an image acquisition head 3, a blue laser light source 5 and a white light LED 6 are provided on the surface of the rod 7. The motor 1 is preferably a servo motor.

[0024] In the preferred embodiment, a main control device is also provided, which has a memory and / or a wireless transmission device. The present invention adopts two data transmission methods, one is to store in the memory and read by the computer after scanning, and the other is to transmit the image to the computer through a wireless transmission device, such as Bluetooth or WiFi.

[0025] The main control device is electrically connected to the motor 1, the image acquisition head 3, the blue laser light source 5 and the white light LED 6;

[0026] A power supply is also provided for supplying power. The main control device preferably adopts an STM32F series chip. The power supply adopts a battery or a DC power supply.

[0027] The preferred solution is Figure 1 In the figure, a co-convergent laser light source 4 is also provided on the rod body 7 along the length direction. The co-convergent laser light source 4 is an array of laser light sources, which are columns along the length direction and rows along the circumferential direction. At least multiple lasers in the same row are focused on the same focal point.

[0028] The preferred solution is Figure 1In the figure, there are 2 to 8 columns of co-convergent laser light sources 4 , which are located on the outer wall of the rod body 7 .

[0029] The preferred solution is Figure 1 In the figure, the co-convergent laser light source 4 is symmetrically arranged on both sides of the image acquisition head 3, the blue laser light source 5 and the white light LED 6.

[0030] In a preferred solution, the image acquisition head 3, the blue laser light source 5 and the white light LED 6 are arranged in a row.

[0031] In a preferred solution, the image acquisition head 3 is located in the middle of the rod body 7 .

[0032] In a preferred solution, there are at least two blue laser light sources 5 and at least two white light LEDs 6 , and the blue laser light sources 5 and the white light LEDs 6 are symmetrically arranged near the two ends of the image acquisition head 3 .

[0033] In a preferred solution, the cross section of the glass cover body 2 is a triangle, and the edges of the triangle are chamfered.

[0034] In a preferred solution, an opening is provided on one end surface of the glass cover body 2, a base is provided at the position of the opening, and the motor 1 is fixed on the base.

[0035] When in use, a three-dimensional model of the oral cavity is established, and the oral cavity scanning device scans the image obtained according to the preset position, maps it to the three-dimensional model of the oral cavity according to the rotation angle parameter and position parameter of the motor 1, and then obtains a plane image by projection. The plane image is input into the corresponding recognition model to obtain the abnormal position of the oral cavity. The oral cavity scanning device of the utility model is easy to use and has a good recognition effect.

[0036] Embodiment 2:

[0037] The oral image screening method using an oral scanning device includes the following methods:

[0038] S01, collecting intraoral images using a multi-light source method;

[0039] S02, inputting the images collected by multi-light source mode into the deep learning recognition model;

[0040] S03, output oral abnormal position assessment;

[0041] S04. Based on the output results, a swab is collected from the oral mucosa at an abnormal position in the oral cavity.

[0042] In step S01, an intraoral image is collected by a conventional scanning method using an oral scanning device, and during the collection process, the intraoral image is collected using blue laser and white light to obtain a fluorescent image a and a white light image b. The wavelength of the blue laser is 415nm~530nm.

[0043] In step S02, the fluorescent image a and the white light image b are input into different recognition models A and B respectively, and the output results are input into the decision tree, random forest tree or gradient boosting tree model to output the abnormal position of the oral cavity. Different light source images are identified by different recognition models respectively, and then the results are outputted according to the weights in the decision tree, random forest tree or gradient boosting tree model, and the output results point to the abnormal position of the oral cavity. The swab collection guided by the abnormal position of the oral cavity can greatly improve the hit rate of the collection and detect clues of oral cancer at an early stage, such as Figure 2 as shown in .

[0044] In another optional scheme, in step S01, after the conventional scanning method, it also includes the step of rinsing the mouth with an indicative fluorescent agent, and the indicative fluorescent agent refers to the step of collecting the intraoral image with an oral scanning device, scanning the inside of the oral cavity with a co-convergent laser light source during the collection process, and scanning the laser-excited oral histological image c. Optionally, the indicative fluorescent agent, such as a substance targeting c-Met based on oral squamous cell carcinoma, tongue squamous cell carcinoma, pharyngeal squamous cell carcinoma or laryngeal squamous cell carcinoma, is a cMBP polypeptide fluorescent imaging agent with an amino acid sequence of KSLSRHDHIHHHK. The wavelength of the co-convergent laser light source is and 765~785nm laser light source.

[0045] The oral scanning device is placed in the left and right sides of the oral cavity and the bottom of the tongue respectively, and then the oral scanning device automatically collects images. When the co-convergent laser light source is started, multiple light source units simultaneously focus on one point to stimulate the fluorescent image of the indicative fluorescent agent, and the image acquisition head (3) collects the fluorescent image stimulated by the co-convergent laser light source in a scanning manner.

[0046] In a preferred embodiment, in step S02, the fluorescent image a and the white light image b are input into different recognition models A and B respectively, the oral histology image c is input into the recognition model c, and the output result is input into a decision tree, a random forest tree or a gradient boosting tree model to output the oral abnormality location;

[0047] Before rinsing your mouth, use a swab to collect oral mucosa from abnormal locations in the mouth;

[0048] Combined with the output results of the recognition model c, the sample collected by the swab is selected.

[0049] S1. Use swabs to collect oral mucosa and oral epithelial cell samples;

[0050] S2, HE staining;

[0051] S3, applying the stained oral epithelial cell sample on a glass slide;

[0052] S4, microscopically scanning the oral epithelial cell samples to obtain oral epithelial cell images;

[0053] S5, extracting features of oral epithelial cell images;

[0054] S6, inputting the extracted oral epithelial cell image feature data into the deep learning model CNN;

[0055] The deep learning model CNN adopts a lightweight CNN model with a 5-layer architecture: the first layer is the input layer; the second layer is the convolution and pooling layer that transmits EfficientNet-B0; the third layer is the global average pooling layer; the fourth layer is a fully connected layer with 1024 neurons, preferably, also having a rectified linear unit and a dropout layer rate of 0.5; the fifth layer is an output layer with an s-type activation function; by the optimized lightweight CNN model, the utility model greatly improves the image recognition efficiency.

[0056] The deep learning model CNN uses manually labeled images as training data sets for training, and the current output results are manually labeled and added to the training data sets for expansion;

[0057] The output results of the supplemented training set include positive feedback training sets, which are the results of the deep learning model CNN's correct recognition; and negative feedback training sets, which are the results of the deep learning model CNN's incorrect recognition. By iterating the training data set, the recognition accuracy is further improved.

[0058] S7, output the classification and recognition results of oral epithelial cells, such as Figure 2 as shown in .

[0059] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limitations of the present invention. The embodiments and features in the embodiments of the present application can be arbitrarily combined with each other without conflict. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. An oral scanning device, characterized in that: The invention comprises a glass cover body (2), wherein the glass cover body (2) has a cavity therein, the glass cover body (2) extends in a length direction, a motor (1) is provided on an end surface of one end of the glass cover body (2), an output shaft of the motor (1) is connected to a rod body (7), the rod body (7) is arranged along the length direction of the glass cover body (2), and an image acquisition head (3), a blue laser light source (5) and a white light LED (6) are provided on the surface of the rod body (7).

2. The oral scanning device according to claim 1, characterized in that: A main control device is also provided, wherein a memory and / or a wireless transmission device is provided in the main control device; The main control device is electrically connected to the motor (1), the image acquisition head (3), the blue laser light source (5) and the white light LED (6); A power supply is also provided for supplying power.

3. An oral scanning device according to claim 1, characterized in that: A co-focused laser light source (4) is also provided on the rod body (7) along the length direction. The co-focused laser light source (4) is an array of laser light sources, which are arranged in columns along the length direction and in rows along the circumferential direction. At least a plurality of lasers in the same row are focused on the same focal point.

4. An oral scanning device according to claim 3, characterized in that: The co-convergent laser light sources (4) are in 2 to 8 rows, and the co-convergent laser light sources (4) are located on the outer wall of the rod body (7).

5. An oral scanning device according to claim 4, characterized in that: The co-convergent laser light sources (4) are symmetrically arranged on both sides of the image acquisition head (3), the blue laser light source (5) and the white light LED (6).

6. An oral scanning device according to claim 4, characterized in that: The image acquisition head (3), the blue laser light source (5) and the white light LED (6) are arranged in a row.

7. An oral scanning device according to claim 6, characterized in that: The image acquisition head (3) is located in the middle of the rod body (7); There are at least two blue laser light sources (5) and at least two white light LEDs (6), and the blue laser light sources (5) and the white light LEDs (6) are symmetrically arranged near two ends of the image acquisition head (3).

8. An oral scanning device according to claim 2, characterized in that: A heating device is also provided in the glass cover (2), and the heating device is electrically connected to the main control device so as to heat the oral scanning device to human body temperature, so as to prevent water vapor from being generated in the glass cover (2) and affecting the field of vision.

9. The oral scanning device according to claim 1, characterized in that: The cross section of the glass cover body (2) is a triangle, and the edges of the triangle are chamfered.

10. The oral scanning device according to claim 1, characterized in that: An opening is provided on one end surface of the glass cover body (2), a base is provided at the position of the opening, and the motor (1) is fixedly mounted on the base.

Images