Tumor image analysis method and system based on artificial intelligence, and storage medium

Abstract

The invention relates to the technical field of image processing, and discloses a tumor image analysis method and system based on artificial intelligence, and a storage medium, and the method comprises the steps: obtaining a plurality of continuous first images shot by first medical imaging equipment, carrying out the preprocessing of the first images, and obtaining a comparison index of the first images after dimension reduction, when the contrast ratio or brightness of the images meets a preset threshold value, the first artificial intelligence model detects the positions of similar tumors in the first images, marks the positions with a first shape and outputs the positions, further judges the switching speed among the multiple first images, and performs secondary processing on the images in the first shape based on the switching speed to obtain a second image; and judging whether the image in the first shape contains a tumor or not, classifying the image in the first shape by using a second artificial intelligence model, and displaying a classification result in an annotation form through a display unit. The efficiency and accuracy of tumor detection are further improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an artificial intelligence-based tumor image analysis method, system and storage medium. Background Art

[0002] In the process of tumor diagnosis and treatment, accurate analysis of tumor images is crucial for early detection of the disease, condition assessment, and formulation of treatment plans. Traditional tumor image analysis methods mainly rely on doctors' experience and naked eye observation, which is not only time-consuming and labor-intensive, but also difficult to ensure accuracy and consistency.

[0003] The Chinese patent application with publication number CN111311578A provides an artificial intelligence-based object classification method, device, computer-readable storage medium and computer equipment, including: obtaining an image to be processed, segmenting a target detection object image of a target detection object from the image to be processed; inputting the target detection object image into a feature object prediction model to obtain a feature object segmentation map of the feature object in the target detection object image; obtaining quantitative feature information of the target detection object based on the target detection object image and the feature object segmentation map; classifying the target detection object image based on the quantitative feature information to obtain the category information of the target detection object in the image to be processed. Although this application can improve the accuracy of image detection, it does not consider the detection efficiency of tumor images. Summary of the invention

[0004] In order to solve the technical problems in the above-mentioned prior art, the present application provides a tumor image analysis method, system and storage medium based on artificial intelligence to realize automatic analysis of tumor images, thereby further improving the efficiency and accuracy of tumor detection.

[0005] In a first aspect, the present application provides a tumor image analysis method based on artificial intelligence, the method comprising:

[0006] Step S1: acquiring a plurality of consecutive first images taken by a first medical imaging device, preprocessing the first images, and acquiring a contrast index of the first images after dimensionality reduction;

[0007] Step S2: inputting the contrast index of the first image after dimensionality reduction into the first artificial intelligence model, the first artificial intelligence model determines whether the contrast and brightness of the first image are greater than a preset threshold based on the contrast index of the first image after dimensionality reduction, if so, the first artificial intelligence model detects a tumor-like position in the first image, marks it with a first shape and outputs it, otherwise further processes the first image;

[0008] Step S3: determining a switching speed between the plurality of first images, and performing secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor, and marking the position and size of the tumor with a second shape when the image in the first shape contains a tumor, wherein the second shape and the first shape use different border colors;

[0009] Step S4: using a second artificial intelligence model to classify the image within the first shape, and outputting a classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models;

[0010] Step S5: displaying the classification result in the form of annotation on one side of the first shape.

[0011] In a second aspect, the present application provides a tumor image analysis system based on artificial intelligence, the system comprising:

[0012] an acquisition unit, configured to acquire a plurality of consecutive first images taken by a first medical imaging device, preprocess the first images, and acquire a contrast index of the first images after dimensionality reduction;

[0013] a first detection unit, configured to input the contrast index of the first image after dimensionality reduction into a first artificial intelligence model, wherein the first artificial intelligence model determines whether the contrast and brightness of the first image are both greater than a preset threshold value based on the contrast index of the first image after dimensionality reduction, and if so, the first artificial intelligence model detects a tumor-like position in the first image, marks it with a first shape and outputs it, otherwise, further processes the first image;

[0014] a second detection unit, configured to determine a switching speed between the plurality of first images, and perform secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor, and mark the position and size of the tumor with a second shape when the image in the first shape contains a tumor, wherein the second shape and the first shape use different border colors;

[0015] a classification unit, configured to classify the image within the first shape using a second artificial intelligence model and output a classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models;

[0016] The display unit is used to display the classification result in the form of annotation on one side of the first shape.

[0017] The third aspect of the present application provides a computer-readable storage medium, which stores instructions, and when the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned tumor image analysis method based on artificial intelligence.

[0018] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0019] The technical solution provided by the present application is to obtain a plurality of consecutive first images taken by a first medical imaging device, pre-process the first images, and obtain the contrast index of the first image after dimensionality reduction to reduce the amount of data calculation; when the contrast or brightness of the image meets the preset threshold, the first artificial intelligence model detects the position of the tumor in the first image, marks it with a first shape and outputs it, and avoids misjudgment caused by image quality problems through threshold judgment, thereby improving the detection accuracy; the switching speed between the plurality of first images is also judged, and the image in the first shape is secondary processed based on the switching speed to judge whether the image in the first shape contains a tumor, and the position and size of the tumor are marked with a second shape when the image in the first shape contains a tumor, and the detection efficiency and accuracy of the tumor are balanced by judging the switching speed; the second artificial intelligence model is used to classify the images in the first shape to assist doctors in making more accurate diagnostic decisions and reduce the occurrence of misdiagnosis and missed diagnosis; and the classification results are displayed on one side of the first shape in the form of annotations through the display unit. The present application further improves the efficiency and accuracy of tumor detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0021] Figure 1 This is a schematic diagram of an embodiment of a tumor image analysis method based on artificial intelligence in an embodiment of the present application;

[0022] Figure 2 This is a schematic diagram of an embodiment of an artificial intelligence-based tumor image analysis system in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The embodiments of the present application provide a tumor image analysis method, system and storage medium based on artificial intelligence. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] Embodiment 1:

[0025] See also Figure 1 In the embodiments of the present application, an embodiment of the tumor image analysis method based on artificial intelligence includes:

[0026] Step S1: obtaining a plurality of consecutive first images taken by a first medical imaging device, preprocessing the first images, and obtaining a comparison index of the first images after dimensionality reduction.

[0027] Specifically, a plurality of continuous first images are obtained from a first medical imaging device (such as a CT scanner, an MRI device, etc.), where the first image is raw data obtained by scanning a specific part of the patient's body, such as a series of lung cross-sectional images obtained by CT scanning the lungs. By preprocessing the first image, the processing speed of the subsequent artificial intelligence model is accelerated, thereby improving the efficiency of tumor detection. The preprocessing of the first image will be described in detail below.

[0028] Step S2: Input the contrast index of the first image after dimensionality reduction into the first artificial intelligence model. The first artificial intelligence model determines whether the contrast and brightness of the first image are greater than a preset threshold based on the contrast index of the first image after dimensionality reduction. If so, the first artificial intelligence model detects the tumor-like position in the first image, marks it with a first shape and outputs it; otherwise, the first image is further processed.

[0029] Specifically, the first artificial intelligence model analyzes the input contrast index according to the patterns and rules learned during training, and determines whether the contrast and brightness of the first image are greater than the preset threshold. For example, the artificial intelligence model may determine whether the image has sufficient clarity and brightness for accurate tumor detection by comparing the contrast index with the threshold. If the contrast and brightness of the image meet the requirements, the first artificial intelligence model uses the tumor feature pattern it has learned to detect the location of a tumor-like tumor in the image. By analyzing the pixel value distribution, texture features, shape features and other information in the first image, the area that may be a tumor is identified. Then the position is marked with a first shape (such as a rectangular frame or an elliptical frame), and the marked image is output. For example, in a lung CT image, the model may identify an area in the lung tissue with abnormal density and a shape that conforms to the characteristics of the tumor, and circle it with a rectangular frame or an elliptical frame. For the first image that meets the conditions, the first artificial intelligence model can quickly locate the possible tumor location, provide an important reference area for further confirmation and analysis, reduce the search range of subsequent processing, and improve analysis efficiency. If the contrast or brightness of the image does not meet the preset threshold, the contrast and brightness of the image can be adjusted again and the contrast index can be recalculated and input into the model for judgment again, or it can be directly discarded if it cannot meet the requirements through processing. In the above steps, the threshold judgment is used to ensure that only images with qualified quality (appropriate contrast and brightness) enter the tumor detection link, avoiding misjudgment due to image quality problems and improving detection accuracy.

[0030] Step S3: Determine the switching speed between the multiple first images, and perform secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor. If the image in the first shape contains a tumor, use a second shape to mark the position and size of the tumor. The second shape and the first shape use different border colors.

[0031] Specifically, if the image switching speed is fast, it indicates that the image may be browsed quickly or dynamically scanned (such as observing the changes of organs in physiological activities). At this time, a faster response to the test results may be required. In this case, secondary processing may not be performed, thereby reducing the amount of calculation to quickly respond to the test results; if the image switching speed is slow or general, a more complex and accurate secondary processing algorithm can be used to make full use of time to improve the detection accuracy. For example, in large-scale rapid screening, not performing secondary processing can improve efficiency; after performing secondary processing, the detection accuracy will be improved. By judging the switching speed, a balance is made between the detection efficiency and accuracy of the tumor. The so-called secondary processing is to further confirm whether the image in the first shape is a tumor through more in-depth feature extraction and analysis, reduce the false positive results that may be generated by the initial detection (i.e., misjudging the non-tumor area as a tumor), improve the accuracy of tumor judgment, and provide patients with a more reliable diagnosis basis. At the same time, if the image in the first shape is judged to contain a tumor, the position and size of the tumor are marked with a second shape (such as a red circular frame, which is different from the border color of the first shape, such as the first shape is a blue rectangular frame) to make the tumor area more eye-catching. The visualization of the image is enhanced through marks and annotations of different shapes and colors, allowing doctors to more intuitively observe tumor-related information in the image and quickly obtain key diagnostic points.

[0032] Step S4: Use the second artificial intelligence model to classify the image within the first shape and output the classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models.

[0033] Specifically, the image within the first shape is input into the second artificial intelligence model. The second artificial intelligence model comprehensively analyzes the image based on its pre-trained tumor classification knowledge. The second artificial intelligence model learns the characteristic patterns of different types of tumors in the image, including the morphological characteristics, texture characteristics, grayscale characteristics and relationship characteristics with surrounding tissues of the tumor. According to the degree of matching between the image features and the patterns learned by the model, the second artificial intelligence model outputs a classification result, such as classifying the tumor as a benign tumor, a malignant tumor or an uncertain type. For example, in a breast MRI image, if the tumor area presents a regular shape, smooth edges, uniform texture and clear boundaries with surrounding tissues, the artificial intelligence model may classify it as a benign tumor; conversely, if the tumor has an irregular shape, fuzzy edges, complex texture and signs of invading surrounding tissues, it may be classified as a malignant tumor. The classification results can assist doctors in making more accurate diagnostic decisions, reduce the occurrence of misdiagnosis and missed diagnosis, and help develop personalized treatment plans. For example, for benign tumors, regular observation or conservative treatment may be sufficient, while malignant tumors require more active treatments such as surgery, chemotherapy or radiotherapy.

[0034] Step S5: Display the classification result in the form of annotation on one side of the first shape.

[0035] Specifically, the classification result is displayed in the form of annotations on one side of the first shape (such as a rectangular frame), clearly informing the doctor or operator of the tumor classification of the image in the region. For example, on the displayed lung CT image, next to the suspected tumor area marked by the rectangular frame, annotation information such as "high probability of benign tumor - lipoma" is displayed. At the same time, clear markings and annotations reduce the risk of misjudgment caused by visual fatigue or negligence of doctors, and improve the accuracy of diagnosis.

[0036] Through the coordination between the above steps, the present application further improves the efficiency and accuracy of tumor detection.

[0037] In a specific embodiment, in step S1, preprocessing the first image includes:

[0038] Image content is extracted from the first image as the first image, a blur operation is performed on the first image, and a second image corresponding to the first image is acquired.

[0039] Based on the first image and the second image, a contrast index between the first image and the second image is calculated, a principal component analysis is performed on the contrast index, the dimension of the contrast index is reduced, and the contrast index of the first image after the dimension reduction is obtained, the contrast index includes a first contrast index, a second contrast index, a third contrast index and a fourth contrast index.

[0040] Specifically, by obtaining the contrast index after dimensionality reduction through the above preprocessing operation, it is possible to more accurately determine whether the image has sufficient clarity and contrast for tumor detection, and how to identify the tumor area from the image features. This helps to improve the training effect of the model, so that it can more accurately detect the tumor location in practical applications, reduce the occurrence of misjudgment (such as misjudging normal tissue as tumor or vice versa), and thus improve the accuracy of the entire tumor image analysis method.

[0041] In a specific embodiment, step S3 further includes:

[0042] Calculate the average pixel difference between two adjacent first images, and calculate the switching speed based on the average pixel difference, and obtain the first processing time corresponding to the switching speed based on the preset image switching schedule, wherein the faster the switching speed of the image recorded in the image switching schedule, the shorter the corresponding first processing time.

[0043] The time from the generation of the first image to the output of the classification result is calculated as the second processing time, and the second processing time is compared with the first processing time. If the second processing time is greater than the first processing time, the computational complexity of the first image is reduced; otherwise, the switching speed and processing time of the first image are continuously monitored.

[0044] Specifically, by dynamically calculating and comparing the image switching speed and processing time, the image operation complexity is adjusted according to the results. While ensuring the accuracy of tumor detection, it adapts to different image switching scenarios and improves the overall analysis efficiency.

[0045] In a specific embodiment, in step S3, the image in the first shape is subjected to secondary processing, including:

[0046] The pixel values ​​of the image within the first shape are obtained and input into the third artificial intelligence model, which outputs a judgment result. When the judgment result is a tumor, the size and position of the tumor are marked with the second shape. Otherwise, the first shape is marked as non-tumor. The third artificial intelligence model has learned a large amount of image data features containing tumors and other substances during the training stage.

[0047] Specifically, this step can further improve the accuracy of tumor detection, effectively distinguish between tumor and non-tumor areas, and improve the accuracy of the entire tumor image analysis method by utilizing the third artificial intelligence model to perform secondary processing and judgment on the image within the first shape based on pixel values.

[0048] In a specific embodiment, the data input to the third artificial intelligence model also includes:

[0049] In view of the continuity of image shooting, the position of the tumor between adjacent images has a certain correlation and movement pattern. Based on the image taken before the first image and the determined second shape of the tumor and preset rules, the first shape of the first image can be expanded, and the shape after the expansion of the first shape is obtained as an alternative. The preset rules are set according to the resolution of the image within the first shape and the time interval between adjacent images.

[0050] Pixel values ​​of the image corresponding to the positions of the alternative items are extracted from the first image and input into the third artificial intelligence model.

[0051] Specifically, the first shape is expanded to obtain alternative options based on the image continuity law, and the corresponding pixel values ​​are extracted and input into the third artificial intelligence model, which helps to make full use of the correlation information between images, expand the range of tumor detection, improve the accuracy of tumor identification, and reduce missed detections due to tumor movement or morphological changes.

[0052] In a specific embodiment, reducing the computational complexity of the first image includes:

[0053] Skip the secondary processing of the image within the first shape or display the first shape marking result first, and then display the remaining results except the first shape marking result.

[0054] In a specific embodiment, the first artificial intelligence model refers to:

[0055] The first artificial intelligence model is used to determine a first shape of a tumor in a first image, using historical image data acquired by a second medical imaging device as learning data, wherein the historical images include tumor morphology, size, and location under different conditions, and also include images without tumors.

[0056] In a specific embodiment, the second artificial intelligence model refers to:

[0057] The second artificial intelligence model is used to classify images within the first shape in the first image, using partial image data in the historical image data as learning data, wherein the partial image data refers to image data with marked tumor types in the historical images.

[0058] Embodiment 2:

[0059] See also Figure 2 In the embodiments of the present application, an embodiment of the tumor image analysis system based on artificial intelligence includes the following modules:

[0060] The acquisition unit is used to acquire a plurality of continuous first images taken by a first medical imaging device, preprocess the first images, and acquire a comparison index of the first images after dimensionality reduction.

[0061] The first detection unit is used to input the contrast index of the first image after dimensionality reduction into the first artificial intelligence model. The first artificial intelligence model determines whether the contrast and brightness of the first image are greater than a preset threshold based on the contrast index of the first image after dimensionality reduction. If so, the first artificial intelligence model detects a tumor-like position in the first image, marks it with a first shape and outputs it; otherwise, the first image is further processed.

[0062] The second detection unit is used to determine the switching speed between the multiple first images, and perform secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor. When the image in the first shape contains a tumor, the position and size of the tumor are marked with a second shape, and the border color used by the second shape and the first shape is different.

[0063] A classification unit is used to classify the image within the first shape using a second artificial intelligence model and output a classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models.

[0064] The display unit is used to display the classification result in the form of annotation on one side of the first shape.

[0065] Through the collaborative efforts of the above components, this application further improves the efficiency and accuracy of tumor detection.

[0066] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the above-mentioned artificial intelligence-based tumor image analysis method.

[0067] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0068] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art 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 a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0069] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A tumor image analysis method based on artificial intelligence, characterized in that: The method comprises: Step S1: acquiring a plurality of consecutive first images taken by a first medical imaging device, preprocessing the first images, and acquiring a contrast index of the first images after dimensionality reduction; Step S2: inputting the contrast index of the first image after dimensionality reduction into the first artificial intelligence model, the first artificial intelligence model determines whether the contrast and brightness of the first image are greater than a preset threshold based on the contrast index of the first image after dimensionality reduction, if so, the first artificial intelligence model detects a tumor-like position in the first image, marks it with a first shape and outputs it, otherwise further processes the first image; Step S3: determining a switching speed between the plurality of first images, and performing secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor, and marking the position and size of the tumor with a second shape when the image in the first shape contains a tumor, wherein the second shape and the first shape use different border colors; Step S4: using a second artificial intelligence model to classify the image within the first shape, and outputting a classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models; Step S5: displaying the classification result in the form of annotation on one side of the first shape.

2. The method according to claim 1, characterized in that In the step S1, preprocessing the first image includes: Extracting image content from the first image as a first image, performing a blur operation on the first image, and acquiring a second image corresponding to the first image; Based on the first image and the second image, a contrast index between the first image and the second image is calculated, a principal component analysis is performed on the contrast index to reduce the dimension of the contrast index, and the contrast index of the first image after the dimension reduction is obtained, the contrast index including a first contrast index, a second contrast index, a third contrast index and a fourth contrast index.

3. The method according to claim 1, characterized in that The step S3 further comprises: Calculating an average pixel difference between two adjacent first images, and calculating a switching speed based on the average pixel difference, and acquiring a first processing time corresponding to the switching speed based on a preset image switching schedule, wherein the faster the image switching speed recorded in the image switching schedule is, the shorter the corresponding first processing time is; Calculate the time from the generation of the first image to the output of the classification result as the second processing time, compare the second processing time with the first processing time, if the second processing time is greater than the first processing time, reduce the computational complexity of the first image, otherwise continue to monitor the switching speed and processing time of the first image.

4. The method according to claim 3, characterized in that In the step S3, the image in the first shape is subjected to secondary processing, including: The pixel values ​​of the image within the first shape are obtained and input into the third artificial intelligence model, the third artificial intelligence model outputs a judgment result, and when the judgment result is a tumor, the size and position of the tumor are marked with a second shape, otherwise, the first shape is marked as non-tumor, wherein the third artificial intelligence model has learned a large amount of image data features containing tumors and other substances during the training stage.

5. The method according to claim 4, characterized in that The data input to the third AI model also includes: Based on an image taken before the first image and a second shape of the tumor determined and a preset rule, expanding the first shape of the first image, and obtaining a shape after the expansion of the first shape as an alternative, wherein the preset rule is set according to a resolution of the image within the first shape and a time interval between adjacent images; Pixel values ​​of an image corresponding to the position of the alternative items are extracted from the first image and input into a third artificial intelligence model.

6. The method according to claim 3, characterized in that Reducing the computational complexity of the first image includes: Skip the secondary processing of the image within the first shape or display the first shape marking result first, and then display the remaining results except the first shape marking result.

7. The method according to claim 1, characterized in that The first artificial intelligence model refers to: The first artificial intelligence model is used to determine a first shape of a tumor in the first image, using historical image data acquired by a second medical imaging device as learning data, wherein the historical images include tumor morphology, size, and location under different conditions, and also include images without tumors.

8. The method according to claim 7, characterized in that The second artificial intelligence model refers to: The second artificial intelligence model is used to classify images within the first shape in the first image, using part of the image data in the historical image data as learning data, wherein the part of the image data refers to image data in the historical images with marked tumor types.

9. An artificial intelligence-based tumor image analysis system, used to implement the method according to any one of claims 1 to 8, characterized in that: The system includes the following modules: an acquisition unit, configured to acquire a plurality of consecutive first images taken by a first medical imaging device, preprocess the first images, and acquire a contrast index of the first images after dimensionality reduction; a first detection unit, configured to input the contrast index of the first image after dimensionality reduction into a first artificial intelligence model, wherein the first artificial intelligence model determines whether the contrast and brightness of the first image are both greater than a preset threshold value based on the contrast index of the first image after dimensionality reduction, and if so, the first artificial intelligence model detects a tumor-like position in the first image, marks it with a first shape and outputs it, otherwise, further processes the first image; a second detection unit, configured to determine a switching speed between the plurality of first images, and perform secondary processing on the image in the first shape based on the switching speed to determine whether the image in the first shape contains a tumor, and mark the position and size of the tumor with a second shape when the image in the first shape contains a tumor, wherein the second shape and the first shape use different border colors; a classification unit, configured to classify the image within the first shape using a second artificial intelligence model and output a classification result, wherein the first artificial intelligence model and the second artificial intelligence model are different artificial intelligence models; The display unit is used to display the classification result in the form of annotation on one side of the first shape.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the artificial intelligence-based tumor image analysis method as described in any one of claims 1 to 8 is implemented.

Images