WO2023178972A1

WO2023178972A1 - Intelligent medical film reading method, apparatus, and device, and storage medium

Info

Publication number: WO2023178972A1
Application number: PCT/CN2022/121727
Authority: WO
Inventors: 叶结保
Original assignee: 康键信息技术（深圳）有限公司
Priority date: 2022-03-23
Filing date: 2022-09-27
Publication date: 2023-09-28
Also published as: CN114628010A

Abstract

The present application relates to the technical field of artificial intelligence. Disclosed are an intelligent medical film reading method, apparatus, and device, and a storage medium, for use in increasing image types of artificial intelligence medical film reading. The intelligent medical film reading method comprises: acquiring a target image; calling a preset medical film reading model to perform image type detection on the target image to obtain a target image type; if the target image type is a type I image, performing lesion identification on the target image, and if a lesion area is present in the target image, marking the lesion area in the target image and sending the marked image to a medical care terminal; if the target image type is a type II image, preprocessing the target image to obtain a preprocessed type II image; and performing lesion identification on the preprocessed type II image, and if a lesion area is present in the target image, marking the lesion area in the target image and sending the marked image to the medical care terminal.

Description

Intelligent medical reading methods, devices, equipment and storage media

This application claims priority to the Chinese patent application submitted to the China Patent Office on March 23, 2022, with the application number 202210289590.9 and the invention title "Intelligent Medical Reading Method, Device, Equipment and Storage Medium", the entire content of which is incorporated by reference. incorporated in this application.

Technical field

This application relates to the field of artificial intelligence technology, and in particular to an intelligent medical reading method, device, equipment and storage medium.

Background technique

With economic development and social progress, people's quality of life has improved, and various diseases have gradually become younger. At the same time, with the convenience of medical treatment, many people go to the hospital to see a doctor. Because some diseases require the collection of images of the internal tissues or body surface of the human body through professional medical equipment. Lesion identification generates a large number of internal tissue images or body surface images of the human body. A large number of internal tissue images or body surface images of the human body require manual reading by hospital doctors, which results in a heavy workload of manual reading and can easily lead to misdiagnosis.

With the development of artificial intelligence (Artificial Intelligence, AI) technology, more and more artificial intelligence technologies have been practically applied in various fields. Existing artificial intelligence medical reading can identify lesions on internal tissue images or body surface images of the human body collected by professional medical equipment, thereby obtaining the lesion area.

The inventor realized that the existing artificial intelligence medical reading only performs lesion identification on the internal tissue images or body surface images of the human body collected by professional medical equipment, and the types of images that can be recognized are fewer.

Contents of the invention

This application provides an intelligent medical reading method, device, equipment and storage medium for increasing the types of images for artificial intelligence medical reading.

The first aspect of this application provides an intelligent medical reading method, including: acquiring a target image, wherein the target image is a first-class image or a second-class image to be identified, and the first-class image is collected by professional medical equipment. In vivo images or body surface images, the second type of images are body surface images collected by non-professional medical equipment; call the preset medical reading model to detect the image type of the target image to obtain the target image type; if the target If the image type is the above-mentioned type of image, then perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical care terminal; If the target image type is the second type image, the target image is preprocessed to obtain a preprocessed second type image, wherein the preprocessed second type image is the preset medical reading image. Images that the model can perform lesion identification on; perform lesion identification on the pre-processed second-class images, and if there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical terminal.

The second aspect of this application provides an intelligent medical reading device, including a memory and at least one processor, instructions are stored in the memory, and the memory and the at least one processor are interconnected through lines; the at least one The processor calls the instructions in the memory, and when the processor executes the computer-readable instructions, the following steps are implemented: obtaining a target image, where the target image is a first-class image or a second-class image to be recognized, The first type of images are in-vivo images or body surface images collected by professional medical equipment, and the second type of images are body surface images collected by non-professional medical equipment; the preset medical reading model is called to perform image type analysis on the target image. Detect to obtain the target image type; if the target image type is the first type of image, perform lesion identification on the target image, and if there is a lesion area in the target image, mark the lesion area in the target image And send the marked image to the medical terminal; if the target image type is the second type image, preprocess the target image to obtain a preprocessed second type image, wherein the preprocessed second type image The class image is an image that can be used for lesion identification by the preset medical reading model; perform lesion identification on the preprocessed second class image, and if there is a lesion area in the target image, the target image is The lesion area is marked and the marked image is sent to the medical terminal.

A third aspect of the present application provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is run on a computer, it causes the computer to perform the following steps: obtain a target image, wherein , the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a body-surface image collected by non-professional medical equipment. ; Call a preset medical reading model to perform image type detection on the target image to obtain the target image type; if the target image type is the first type of image, perform lesion identification on the target image. If the target image If there is a lesion area in the target image, the lesion area is marked in the target image and the marked image is sent to the medical care terminal; if the target image type is the second type image, the target image is preprocessed , obtain a pre-processed second-class image, wherein the pre-processed second-class image is an image that can be used for lesion identification by the preset medical reading model; perform lesion identification on the pre-processed second-class image, if If a lesion area exists in the target image, the lesion area is marked in the target image and the marked image is sent to the medical care terminal.

The fourth aspect of this application provides an intelligent medical reading device, including: an acquisition module for acquiring a target image, wherein the target image is a first-class image or a second-class image to be recognized, and the first-class image is In vivo images or body surface images collected by professional medical equipment. The second type of images are body surface images collected by non-professional medical equipment; a detection module is used to call a preset medical reading model to perform image type detection on the target image. , obtain the target image type; the first recognition sending module is used to perform lesion identification on the target image if the target image type is the first type of image, and if there is a lesion area in the target image, Mark the lesion area in the image and send the marked image to the medical terminal; a preprocessing module is used to preprocess the target image if the target image type is the second type image to obtain the preprocessed image Class II images, wherein the pre-processed Class II images are images that can be used for lesion identification by the preset medical reading model; the second recognition sending module is used to perform processing on the pre-processed Class II images. Lesion identification, if there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical care terminal.

In the technical solution provided by this application, a target image is obtained, in which the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a non-professional image. Body surface images collected by medical equipment; call the preset medical reading model to detect the image type of the target image and obtain the target image type; if the target image type is a type of image, perform lesion identification on the target image. If a lesion area exists, mark the lesion area in the target image and send the marked image to the medical terminal; if the target image type is a Class II image, preprocess the target image to obtain a preprocessed Class II image. Among them, the pre-processed second-class images are images that can be used for lesion identification by the preset medical interpretation model; the pre-processed second-class images are used for lesion identification. If there is a lesion area in the target image, the lesion area is identified in the target image. Mark and send the marked image to the medical terminal. In the embodiment of the present application, the target image is acquired and the image type is detected. If the target image is an in-vivo image or a body surface image collected by professional medical equipment, the target image is directly identified as a lesion. If the target image is a non-professional For body surface images collected by medical equipment, the target image is preprocessed. The preprocessed target image is an image that can be used for lesion identification by the preset medical reading model, and then the preprocessed target image is used for lesion identification. If If there is a lesion area in the target image, the lesion area will be marked in the target image and the marked image will be sent to the medical care terminal, increasing the types of images for artificial intelligence medical interpretation.

Description of the drawings

Figure 1 is a schematic diagram of an embodiment of the intelligent medical reading method in the embodiment of the present application;

Figure 2 is a schematic diagram of another embodiment of the intelligent medical reading method in the embodiment of the present application;

Figure 3 is a schematic diagram of an embodiment of the intelligent medical reading device in the embodiment of the present application;

Figure 4 is a schematic diagram of another embodiment of the intelligent medical reading device in the embodiment of the present application;

Figure 5 is a schematic diagram of an embodiment of the intelligent medical reading device in the embodiment of the present application.

Detailed ways

The terms "first", "second", "third", "fourth", etc. (if present) in the description and claims of this application and the above-mentioned drawings are used to distinguish similar objects without necessarily using Used to describe a specific order or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances so that the embodiments described herein can be practiced in sequences other than those illustrated or described herein. In addition, the terms "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusions, e.g., processes, methods, systems, products, or devices that comprise a series of steps or units and are not necessarily limited to those expressly listed. steps or units, but may include other steps or units not expressly listed or inherent to such processes, methods, products or apparatuses.

The embodiments of this application can obtain and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .

Basic artificial intelligence technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometric technology, speech processing technology, natural language processing technology, and machine learning/deep learning.

For ease of understanding, the specific process of the embodiment of the present application is described below. Please refer to Figure 1. An embodiment of the intelligent medical reading method in the embodiment of the present application includes:

101. Obtain the target image, where the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body surface image collected by professional medical equipment, and the second-class image is a body surface collected by non-professional medical equipment. image;

It can be understood that the execution subject of this application can be an intelligent medical reading device or a terminal, which is not specifically limited here. The embodiment of this application takes an intelligent medical reading device as the execution subject as an example for description.

In this embodiment, the target image can be obtained through a variety of data. For example, the intelligent medical reading device can obtain the target image through the online patient consultation data of the Internet hospital, or obtain the target image through the patient's health cabin physical examination report or health space station physical examination report. Image, or obtain the target image through the graphic and text data of the patient's registration for health lectures.

102. Call the preset medical reading model to detect the image type of the target image and obtain the target image type;

In this embodiment, the spectrum of in-vivo images or body-surface images collected by professional medical equipment is reduced by eliminating light scattering, that is, the noise distribution of in-vivo images or body-surface images collected by professional medical equipment is approximated as Poisson Distribution, most of the in vivo images or body surface images collected by professional medical equipment are single-channel grayscale images. All the information contained in the in vivo images or body surface images collected by professional medical equipment has potential utilization value. For example, human tissues have high Similarity, a slight change in the in-vivo images or surface images collected by professional medical equipment may represent diseased tissue. Due to the existence of light scattering, the spectrum of body surface images collected by non-professional medical equipment is relatively wide, that is, the noise distribution of the body surface images collected by non-professional medical equipment is approximated as a Gaussian distribution.

103. If the target image type is a type of image, perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical terminal;

In this embodiment, there are usually both target areas and non-target areas in the target image. The non-target areas usually interfere with the recognition of the target area. Therefore, the target area can be segmented and the non-target areas can be filtered out to obtain target area. For lesion identification in target images, the location of pixels in the target image whose pixel values are within a preset range can be estimated as the lesion area through pixel-level identification. The preset range of pixel values can be set based on historical experience. The target image can also be input into a segmentation model trained to convergence, and the lesion area in the target image is determined through this segmentation model.

104. If the target image type is a Class II image, preprocess the target image to obtain a preprocessed Class II image, where the preprocessed Class II image is an image that can be used for lesion identification by the preset medical reading model;

In this embodiment, an intelligent medical reading device is used to preprocess body surface images collected by non-professional medical equipment. Only the preprocessed images can be used to identify lesions through a preset medical reading model. The preprocessing process includes Noise reduction, grayscale and binarization. Noise reduction is used to reduce the Gaussian noise of body surface images collected by non-professional medical equipment into Poisson noise. Grayscale is used to convert body surface images collected by non-professional medical equipment. From a three-channel image to a single-channel image, binarization is used to select target areas and exclude non-target areas.

105. Perform lesion identification on the preprocessed second-class image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical terminal.

In this embodiment, the convolutional neural network can be used to perform lesion identification on the preprocessed Class II images, or the preprocessed Class II images can be identified through a visual transformer. Convolutional neural networks include a variety of neural networks, such as regional convolutional neural network Region-CNN, spatial pyramid pooling convolutional neural network SPP-Net, object detection convolutional neural network Yolo, object detection convolutional neural network SSD, etc. Convolutional neural network.

In the embodiment of the present application, a target image is obtained, where the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a non-professional medical device. Collected body surface images; call the preset medical reading model to detect the image type of the target image, and obtain the target image type; if the target image type is a type of image, perform lesion identification on the target image, and if there are lesions in the target image area, mark the lesion area in the target image and send the marked image to the medical terminal; if the target image type is a Class II image, preprocess the target image to obtain a preprocessed Class II image, where, The pre-processed second-class images are images that can be used for lesion identification by the preset medical interpretation model; perform lesion identification on the pre-processed second-class images, and if there is a lesion area in the target image, mark the lesion area in the target image And the marked images are sent to the medical terminal, increasing the types of images for artificial intelligence medical interpretation.

Please refer to Figure 2. Another embodiment of the intelligent medical reading method in the embodiment of the present application includes:

201. Generate a preset medical reading model based on the first-class images and second-class images of historical patients, where there are lesion areas in the first-class images and second-class images;

Specifically, (1) the intelligent medical reading device acquires first-class images and second-class images of historical patients, in which there are lesion areas in the first-class images and second-class images; (2) the first-class images are extracted to obtain the first-class images. The target lesion area in the image is marked, and the target lesion area is marked to obtain a labeled first-class image; (3) The second-class image is extracted to obtain the target lesion area in the second-class image, and the target lesion area is marked. Obtain labeled Class II images; (4) Carry out model training based on the labeled Class I images and labeled Class II images to generate a preset medical reading model.

For example, an intelligent medical reading device acquires first-class images and second-class images of historical patients. The first-class images are in-vivo images or surface images collected by professional medical equipment. The in-vivo images collected by professional medical equipment can be magnetic resonance MRI images. It can also be an X-ray image. The body surface image collected by professional medical equipment can be an image collected by a medical microscope imaging equipment. The second type of image is a body surface image collected by non-professional medical equipment. The body surface image collected by non-professional medical equipment can be It is a human face image or a human leg image. Among them, there are lesion areas in the first-class images and second-class images of historical patients; the first-class images are extracted to obtain the target lesion areas in the first-class images. If the target If the target lesion area is a lung tumor, mark the lung tumor and obtain a labeled lung tumor image. If the target lesion area is a body surface hemangioma, mark the body surface hemangioma and obtain a labeled body surface hemangioma image. ; Extract the second-class image to obtain the target lesion area in the second-class image. If the target lesion area is facial acne, mark the facial acne and obtain the marked facial acne image. If the target lesion area is strip-shaped, Herpes, mark herpes zoster to obtain a labeled herpes zoster image; perform model training based on the labeled lung tumor image or labeled body surface hemangioma image and the labeled facial acne image or labeled herpes zoster image , generate a preset medical reading model.

202. Obtain the target image, where the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body surface image collected by professional medical equipment, and the second-class image is a body surface collected by non-professional medical equipment. image;

In this embodiment, professional medical equipment includes a variety of equipment, such as magnetic resonance imaging equipment, computed tomography CT imaging equipment, X-ray imaging equipment, ultrasound imaging equipment, positron emission tomography PET imaging equipment, endoscopy Equipment, medical microscope imaging equipment and other professional medical equipment. Non-professional medical equipment includes a variety of equipment, such as mobile phones, cameras, scanners and other non-professional medical equipment.

203. Call the preset medical reading model to detect the image type of the target image and obtain the target image type;

Specifically, (1) the intelligent medical reading device performs image noise analysis on the target image through a preset medical reading model, and obtains the image noise analysis result; (2) If the image noise analysis result shows that the target image has Poisson noise, then The target image type corresponding to the target image is determined as a Class I image; (3) If the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined as a Class II image.

For example, the intelligent medical reading device performs image noise analysis on the target image through a preset medical reading model to obtain the image noise analysis result. If the image noise analysis result shows that the target image has Poisson noise, the target image corresponding to the target image will be The type is determined as a type of image, that is, the target image is an in-vivo image or a body surface image collected by professional medical equipment; if the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined as a type-II image, That is, the target image is a body surface image collected by non-professional medical equipment.

204. If the target image type is a type of image, perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical terminal;

Specifically, (1) if the target image type is a type of image, the intelligent medical reading device will segment the target image to obtain the target image area; (2) if the pixel value of each pixel in the target image area is greater than or equal to the preset pixel value, then it is determined that the lesion area exists in the target image; (3) Mark the lesion area in the target image, and send the marked image to the medical terminal.

After step (2), it also includes: if the pixel value of each pixel point in the target image area is less than the preset pixel value, it is determined that there is no lesion area in the target image, and a reminder message is generated, and the intelligent medical reading device will The reminder information is sent to the medical terminal, and the reminder information is used to instruct the review of the lesion identification result of the target image.

For example, the preset pixel value is 100. If the target image is a lung image, the intelligent medical reading device will segment the lung image to obtain the target image area of the lung image; if the target image area of each pixel in the target image area is If the pixel values are all greater than or equal to 100, it is determined that there is a lesion area in the lung image, the lesion area is marked in the lung image, and the marked lung image is sent to the medical terminal. If the pixel value of each pixel in the target image area is less than 100, it is determined that there is no lesion area in the lung image, and a reminder message is generated. The smart medical reading device sends the reminder message to the medical terminal, and the reminder message is used for instructions. Review the lesion identification results of lung images.

205. If the target image type is a Class II image, preprocess the target image to obtain a preprocessed Class II image, where the preprocessed Class II image is an image that can be used for lesion identification by the preset medical reading model;

Specifically, (1) if the target image type is a Class II image, the intelligent medical reading device performs denoising on the target image to obtain a denoised target image; (2) performs grayscale processing on the denoised target image. , obtain the grayscale target image; (3) Binarize the grayscale target image to obtain the preprocessed second-class image.

For example, if the target image is a skin image, the intelligent medical reading device performs denoising on the skin image to obtain a denoised skin image, that is, converts the Gaussian noise of the skin image into Poisson noise; grayscales the denoised skin image. Grayscale processing is performed to obtain a grayscale skin image, that is, the three-channel skin image is converted into a single-channel skin image; the grayscale skin image is binarized to obtain a preprocessed skin image, where, preprocessing The gray value of the pixels in the skin image is 0 or 255.

206. Perform lesion identification on the preprocessed second-class image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical care terminal.

Specifically, (1) the intelligent medical reading device divides the preprocessed second-class images into image blocks to obtain multiple image blocks; (2) convolves the multiple image blocks one by one according to the specified order to obtain each corresponding The convolution value of the image block; (3) If the convolution value of the image block is greater than or equal to the preset convolution value, it is determined that there is a lesion area in the corresponding image block, and the image block with the lesion area is added to the lesion image block set , the lesion image block set includes multiple lesion image blocks; (4) Mark the lesion area in each of the multiple lesion image blocks, generate a marked target image, and send the marked target image to Medical terminal.

After step (2), it also includes: if the convolution value of each image block is less than the preset convolution value, it is determined that there is no lesion area in the target image, and then a reminder message is generated to review the lesion recognition result of the target image. , and send reminder information to the medical terminal.

For example, the intelligent medical reading device divides the preprocessed Class II image into image blocks to obtain multiple image blocks. The multiple image blocks include No. 1 image block, No. 2 image block, No. 3 image block and No. 4 image block; By convolving multiple image blocks one by one in the order of 1 to 4, the convolution value of image block No. 1 is 0.6, the convolution value of image block No. 2 is 0.4, and the convolution value of image block No. 3 is 0.7 , the convolution value of image block No. 4 is 0.5; if the preset convolution value is 0.5, then the convolution value of image block No. 1 is greater than the preset convolution value, that is, it is determined that there is a lesion area in image block No. 1, and If the No. 1 image block is added to the lesion image block set, the convolution value of the No. 3 image block is greater than the preset convolution value, that is, it is determined that there is a lesion area in the No. 3 image block, and the No. 3 image block is added to the lesion image block set, then The convolution value of image block No. 4 is equal to the preset convolution value, that is, it is determined that there is a lesion area in image block No. 4, and image block No. 4 is added to the lesion image block set. The lesion image block set includes image block No. 1, image block No. 3 Image block and No. 4 image block; mark the lesion area in each of the three lesion image blocks, generate a marked target image, and send the marked target image to the medical terminal. Or by convolving multiple image blocks one by one in the order of 1 to 4, the convolution value of image block No. 1 is 0.3, the convolution value of image block No. 2 is 0.4, and the convolution value of image block No. 3 is 0.3, the convolution value of No. 4 image block is 0.2; if the preset convolution value is 0.5, then the convolution value of each image block is less than the preset convolution value, that is, it is determined that there is no lesion area in the target image, then generate Reminder information to review the lesion identification results of the target image, and send the reminder information to the medical terminal.

The smart medical reading method in the embodiment of the present application is described above, and the smart medical reading device in the embodiment of the present application is described below. Please refer to Figure 3. An embodiment of the smart medical reading device in the embodiment of the present application includes:

The acquisition module 301 is used to acquire a target image, where the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body surface image collected by professional medical equipment, and the second-class image is a non-professional medical equipment. Collected body surface images;

The detection module 302 is used to call a preset medical reading model to detect the image type of the target image and obtain the target image type;

The first recognition sending module 303 is used to perform lesion recognition on the target image if the target image type is a type of image. If there is a lesion area in the target image, mark the lesion area in the target image and store the marked image. Send to medical terminal;

The preprocessing module 304 is used to preprocess the target image to obtain a preprocessed Class II image if the target image type is a Class II image, where the preprocessed Class II image can be processed by a preset medical interpretation model. Images for lesion identification;

The second recognition and sending module 305 is used to perform lesion recognition on the preprocessed Class II image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical terminal.

Please refer to Figure 4. Another embodiment of the intelligent medical reading device in the embodiment of the present application includes:

Optionally, the detection module 302 can also be specifically used to:

Perform image noise analysis on the target image through the preset medical reading model to obtain the image noise analysis results;

If the image noise analysis result shows that the target image contains Poisson noise, then the target image type corresponding to the target image is determined as a type of image;

If the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined to be a Class II image.

Optionally, the first identification sending module 303 includes:

The segmentation unit 3031 is used to segment the target image to obtain the target image area if the target image type is a type of image;

The determination unit 3032 is used to determine that a lesion area exists in the target image if the pixel value of each pixel point in the target image area is greater than or equal to the preset pixel value;

The marking unit 3033 is used to mark the lesion area in the target image, and send the marked image to the medical terminal.

Optionally, the first identification sending module 303 also includes:

The reminder unit 3034 is used to determine that there is no lesion area in the target image if the pixel value of each pixel point in the target image area is less than the preset pixel value, generate reminder information, and send the reminder information to the medical terminal. Used to instruct to review the lesion identification results of the target image.

Optionally, the preprocessing module 304 can also be specifically used to:

If the target image type is a Class II image, perform noise reduction processing on the target image to obtain a denoised target image;

Perform grayscale processing on the noise-reduced target image to obtain a grayscale target image;

Binarize the grayscale target image to obtain a preprocessed second-class image.

Optionally, the second identification sending module 305 can also be specifically used for:

Segment the preprocessed Class II image into image blocks to obtain multiple image blocks;

Convolve multiple image blocks one by one according to the specified order to obtain the corresponding convolution value of each image block;

If the convolution value of the image block is greater than or equal to the preset convolution value, it is determined that there is a lesion area in the corresponding image block, and the image block with the lesion area is added to the lesion image block set. The lesion image block set includes multiple lesion images. piece;

Mark the lesion area in each of the multiple lesion image blocks, generate a marked target image, and send the marked target image to the medical care terminal.

Optionally, the smart medical reading device also includes:

The generation module 306 is used to obtain first-class images and second-class images of historical patients, where there are lesion areas in the first-class images and second-class images;

Extract a class of images to obtain a target lesion area in a class of images, and mark the target lesion area to obtain a labeled class of image;

Extract the second-class image to obtain the target lesion area in the second-class image, and mark the target lesion area to obtain the marked second-class image;

Carry out model training based on the labeled first-class images and labeled second-class images to generate a preset medical reading model.

The above Figures 3 and 4 describe in detail the intelligent medical reading device in the embodiment of the present application from the perspective of modular functional entities. The following describes the intelligent medical reading device in the embodiment of the present application in detail from the perspective of hardware processing.

Figure 5 is a schematic structural diagram of an intelligent medical interpretation device provided by an embodiment of the present application. The intelligent medical interpretation device 500 may vary greatly due to different configurations or performance, and may include one or more central processors (central processors). processing units (CPU) 510 (eg, one or more processors) and memory 520, one or more storage media 530 (eg, one or more mass storage devices) that stores applications 533 or data 532. Among them, the memory 520 and the storage medium 530 may be short-term storage or persistent storage. The program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the intelligent medical reading device 500 . Furthermore, the processor 510 may be configured to communicate with the storage medium 530 and execute a series of instruction operations in the storage medium 530 on the intelligent medical reading device 500 .

The intelligent medical reading device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and/or one or more operating systems 531, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD and more. Those skilled in the art can understand that the structure of the intelligent medical interpretation device shown in Figure 5 does not constitute a limitation on the intelligent medical interpretation device. It may include more or fewer components than shown in the figure, or combine certain components, or Different component arrangements.

This application also provides an intelligent medical reading device. The computer device includes a memory and a processor. Computer-readable instructions are stored in the memory. When the computer-readable instructions are executed by the processor, they cause the processor to execute the above-mentioned embodiments. The steps of the intelligent medical reading method.

This application also provides a computer-readable storage medium. The computer-readable storage medium can be a non-volatile computer-readable storage medium. The computer-readable storage medium can also be a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium, and when the instructions are run on the computer, they cause the computer to execute the steps of the intelligent medical reading method.

Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be described again here.

If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk and other media that can store program code. .

As mentioned above, the above embodiments are only used to illustrate the technical solution of the present application, but not to limit it. Although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still make the foregoing technical solutions. The technical solutions described in each embodiment may be modified, or some of the technical features may be equivalently replaced; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions in each embodiment of the present application.

Claims

An intelligent medical reading method, wherein the intelligent medical reading method includes:

Obtain a target image, wherein the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a non-professional medical device. Collected body surface images;

Call a preset medical reading model to detect the image type of the target image to obtain the target image type;

If the target image type is the first type of image, perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image. to the medical terminal;

If the target image type is the second type image, the target image is preprocessed to obtain a preprocessed second type image, wherein the preprocessed second type image is the preset medical reading image. Images where the model can perform lesion identification;

Lesion identification is performed on the preprocessed second-class image. If there is a lesion area in the target image, the lesion area is marked in the target image and the marked image is sent to the medical care terminal.
The intelligent medical reading method according to claim 1, wherein the calling a preset medical reading model to detect the image type of the target image to obtain the target image type includes:

Perform image noise analysis on the target image through a preset medical reading model to obtain image noise analysis results;

If the image noise analysis result is that Poisson noise exists in the target image, then determine the target image type corresponding to the target image as a type of image;

If the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined to be a Class II image.
The intelligent medical image reading method according to claim 1, wherein if the target image type is the first type of image, then perform lesion identification on the target image, and if there is a lesion area in the target image, then perform lesion identification on the target image. Mark the lesion area in the target image and send the marked image to the medical terminal, including:

If the target image type is the first type of image, segment the target image to obtain the target image area;

If the pixel value of each pixel in the target image area is greater than or equal to the preset pixel value, it is determined that a lesion area exists in the target image;

Mark the lesion area in the target image, and send the marked image to the medical terminal.
The intelligent medical reading method according to claim 3, wherein, after performing segmentation processing on the target image to obtain the target image area if the target image type is the first type of image, it further includes:

If the pixel value of each pixel point in the target image area is less than the preset pixel value, it is determined that there is no lesion area in the target image, and a reminder message is generated, and the reminder message is sent to the medical care terminal. The reminder information is used to instruct the review of the lesion identification result of the target image.
The intelligent medical reading method according to claim 1, wherein if the target image type is the second type image, preprocessing is performed on the target image to obtain a preprocessed second type image, including:

If the target image type is the second type of image, perform noise reduction processing on the target image to obtain a denoised target image;

Perform grayscale processing on the denoised target image to obtain a grayscale target image;

The grayscale target image is binarized to obtain a preprocessed second-class image.
The intelligent medical reading method according to claim 1, wherein the preprocessed second-class image is subjected to lesion identification, and if a lesion area exists in the target image, the lesion area is identified in the target image. Mark and send the marked images to the medical terminal, including:

Perform image block segmentation on the preprocessed Class II image to obtain multiple image blocks;

Convolve the plurality of image blocks one by one according to the specified order to obtain the corresponding convolution value of each image block;

If the convolution value of the image block is greater than or equal to the preset convolution value, it is determined that a lesion area exists in the corresponding image block, and the image block with the lesion area is added to the lesion image block set, and the lesion image block set includes multiple Lesion image block;

Mark the lesion area in each of the plurality of lesion image blocks, generate a marked target image, and send the marked target image to the medical care terminal.
The intelligent medical reading method according to any one of claims 1 to 6, wherein before acquiring the target image, it further includes:

Obtaining first-class images and second-class images of historical patients, wherein there are lesion areas in the first-class images and second-class images;

Extract the first type of image to obtain the target lesion area in the first type of image, and mark the target lesion area to obtain the marked first type of image;

Extract the second class image to obtain the target lesion area in the second class image, and mark the target lesion area to obtain the marked second class image;

Model training is performed based on the labeled first-class images and the labeled second-class images to generate a preset medical reading model.
An intelligent medical reading device, wherein the intelligent medical reading device includes:

Acquisition module, used to acquire a target image, wherein the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body surface image collected by professional medical equipment, and the second-class image Body surface images collected for non-specialized medical devices;

A detection module, used to call a preset medical reading model to detect the image type of the target image to obtain the target image type;

The first identification sending module is used to perform lesion identification on the target image if the target image type is the first type of image, and to perform lesion identification on the lesion area in the target image if there is a lesion area in the target image. Mark and send the marked images to the medical terminal;

A preprocessing module, configured to preprocess the target image to obtain a preprocessed second class image if the target image type is the second class image, wherein the preprocessed second class image is the second class image. The preset medical image reading model can identify images of lesions;

The second recognition and sending module is used to perform lesion recognition on the preprocessed Class II image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image. to the medical terminal.
An intelligent medical interpretation device, wherein the intelligent medical interpretation device includes: a memory and at least one processor, and instructions are stored in the memory;

The at least one processor calls the instructions in the memory, so that the smart medical reading device executes the steps of the smart medical reading method as follows:

Obtain a target image, wherein the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a non-professional medical device. Collected body surface images;

Call a preset medical reading model to detect the image type of the target image to obtain the target image type;

If the target image type is the first type of image, perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image. to the medical terminal;

If the target image type is the second type image, the target image is preprocessed to obtain a preprocessed second type image, wherein the preprocessed second type image is the preset medical reading image. Images where the model can perform lesion identification;

Lesion identification is performed on the preprocessed second-class image. If there is a lesion area in the target image, the lesion area is marked in the target image and the marked image is sent to the medical care terminal.
The intelligent medical interpretation device according to claim 9, wherein the intelligent medical interpretation program is executed by the processor to implement the calling of the preset medical interpretation model to detect the image type of the target image, and obtain target image type, also perform the following steps:

Perform image noise analysis on the target image through a preset medical reading model to obtain image noise analysis results;

If the image noise analysis result is that Poisson noise exists in the target image, then determine the target image type corresponding to the target image as a type of image;

If the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined to be a Class II image.
The intelligent medical interpretation device according to claim 9, wherein the intelligent medical interpretation program is executed by the processor to implement, if the target image type is the first type of image, the target image Perform lesion identification, and if there is a lesion area in the target image, mark the lesion area in the target image and send the marked image to the medical care terminal, the following steps are also performed:

If the target image type is the first type of image, segment the target image to obtain the target image area;

If the pixel value of each pixel in the target image area is greater than or equal to the preset pixel value, it is determined that a lesion area exists in the target image;

Mark the lesion area in the target image, and send the marked image to the medical terminal.
The intelligent medical interpretation device according to claim 11, wherein the intelligent medical interpretation program is executed by the processor to implement if the target image type is the first type of image, then the target image is After the image is segmented and the target image area is obtained, the following steps are also performed:

If the pixel value of each pixel point in the target image area is less than the preset pixel value, it is determined that there is no lesion area in the target image, and a reminder message is generated, and the reminder message is sent to the medical care terminal. The reminder information is used to instruct the review of the lesion identification result of the target image.
The intelligent medical interpretation device according to claim 9, wherein the intelligent medical interpretation program is executed by the processor to implement the step of processing the target image if the target image type is the second type image. When performing preprocessing to obtain the preprocessed Class II image, the following steps are also performed:

If the target image type is the second type of image, perform noise reduction processing on the target image to obtain a denoised target image;

Perform grayscale processing on the denoised target image to obtain a grayscale target image;

The grayscale target image is binarized to obtain a preprocessed second-class image.
The intelligent medical interpretation device according to claim 9, wherein the intelligent medical interpretation program is executed by the processor to implement the lesion identification on the preprocessed second-class image. If the target image contains If a lesion area exists, when marking the lesion area in the target image and sending the marked image to the medical terminal, the following steps are also performed:

Perform image block segmentation on the preprocessed Class II image to obtain multiple image blocks;

Convolve the plurality of image blocks one by one according to the specified order to obtain the corresponding convolution value of each image block;

If the convolution value of the image block is greater than or equal to the preset convolution value, it is determined that a lesion area exists in the corresponding image block, and the image block with the lesion area is added to the lesion image block set, and the lesion image block set includes multiple Lesion image block;

Mark the lesion area in each of the plurality of lesion image blocks, generate a marked target image, and send the marked target image to the medical care terminal.
A computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps of the intelligent medical reading method are implemented as follows:

Obtain a target image, wherein the target image is a first-class image or a second-class image to be identified, the first-class image is an in-vivo image or a body-surface image collected by professional medical equipment, and the second-class image is a non-professional medical device. Collected body surface images;

Call a preset medical reading model to detect the image type of the target image to obtain the target image type;

If the target image type is the first type of image, perform lesion identification on the target image. If there is a lesion area in the target image, mark the lesion area in the target image and send the marked image. to the medical terminal;

If the target image type is the second type image, the target image is preprocessed to obtain a preprocessed second type image, wherein the preprocessed second type image is the preset medical reading image. Images where the model can perform lesion identification;

Lesion identification is performed on the preprocessed second-class image. If there is a lesion area in the target image, the lesion area is marked in the target image and the marked image is sent to the medical care terminal.
The computer-readable storage medium according to claim 15, wherein when the computer program is executed by the processor, the step of calling a preset medical reading model to detect the image type of the target image and obtain the target image type is obtained. , also perform the following steps:

Perform image noise analysis on the target image through a preset medical reading model to obtain image noise analysis results;

If the image noise analysis result is that Poisson noise exists in the target image, then determine the target image type corresponding to the target image as a type of image;

If the image noise analysis result shows that Gaussian noise exists in the target image, the target image type corresponding to the target image is determined to be a Class II image.
The computer-readable storage medium according to claim 15, wherein the computer program is executed by a processor and performs lesion identification on the target image if the target image type is the first type of image. If there is a lesion area in the image, then in the step of marking the lesion area in the target image and sending the marked image to the medical terminal, the following steps are also performed:

If the target image type is the first type of image, segment the target image to obtain the target image area;

If the pixel value of each pixel in the target image area is greater than or equal to the preset pixel value, it is determined that a lesion area exists in the target image;

Mark the lesion area in the target image, and send the marked image to the medical terminal.
The computer-readable storage medium according to claim 17, wherein the computer program is executed by the processor to perform segmentation processing on the target image if the target image type is the first type of image, to obtain After the step of target image area, also perform the following steps:

If the pixel value of each pixel point in the target image area is less than the preset pixel value, it is determined that there is no lesion area in the target image, and a reminder message is generated, and the reminder message is sent to the medical care terminal. The reminder information is used to instruct the review of the lesion identification result of the target image.
The computer-readable storage medium according to claim 15, wherein the computer program is executed by the processor and performs preprocessing on the target image to obtain a pre-processed image if the target image type is the second type image. When processing the second-class image, the following steps are also performed:

If the target image type is the second type of image, perform noise reduction processing on the target image to obtain a denoised target image;

Perform grayscale processing on the denoised target image to obtain a grayscale target image;

The grayscale target image is binarized to obtain a preprocessed second-class image.
The computer-readable storage medium according to claim 15, wherein the computer program is executed by a processor to perform lesion identification on the pre-processed second-class image, and if there is a lesion area in the target image, then When marking the lesion area in the target image and sending the marked image to the medical terminal, the following steps are also performed:

Perform image block segmentation on the preprocessed Class II image to obtain multiple image blocks;

Convolve the plurality of image blocks one by one according to the specified order to obtain the corresponding convolution value of each image block;

If the convolution value of the image block is greater than or equal to the preset convolution value, it is determined that a lesion area exists in the corresponding image block, and the image block with the lesion area is added to the lesion image block set, and the lesion image block set includes multiple Lesion image block;

Mark the lesion area in each of the plurality of lesion image blocks, generate a marked target image, and send the marked target image to the medical care terminal.