Cancer treatment evaluation method and device based on artificial intelligence analysis dynamic image, and storage medium

Abstract

The invention discloses a cancer treatment evaluation method based on artificial intelligence analysis of a dynamic image, a cancer treatment evaluation device based on artificial intelligence analysis of a dynamic image and a storage medium, and the method comprises the steps: firstly receiving a dynamic image, and intercepting at least two to-be-processed images from the dynamic image based on a preset time step length, then, determining an abnormal tissue area, a non-abnormal tissue area and patient identification associated based on a dynamic image, obtaining historical pictures, grouping the historical pictures and the to-be-processed pictures according to the similarity between the historical pictures and the non-abnormal tissue area in the to-be-processed pictures, and according to the similarity between the historical pictures and the non-abnormal tissue area in the to-be-processed pictures, determining the to-be-processed pictures in the to-be-processed pictures; and generating a treatment effect evaluation result in the abnormal tissue area of the historical image and the to-be-processed image. Therefore, the technical problem of poor consistency of analysis results is solved, and the technical effect of improving the consistency of the analysis results is achieved.

Description

Technical Field

[0001] The present application relates to the field of electronic digital data processing, and in particular to a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images, a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, and a storage medium. Background Art

[0002] Dynamic Medical Imaging is an imaging method that uses medical imaging technology to observe the changes in the internal structure, function or physiological process of the human body over time in real time or continuously. Compared with traditional static imaging technology (such as X-rays, conventional CT or MRI single scans), dynamic medical imaging can capture dynamic processes such as organ movement, hemodynamic changes, metabolic activity and drug distribution, thereby providing more comprehensive spatiotemporal information for clinical diagnosis and treatment. Especially in the field of cancer treatment, dynamic imaging technology is widely used in surgical planning, radiotherapy target positioning and staged evaluation of treatment effects. By collecting dynamic imaging data of the lesion area during different treatment cycles (such as dynamic enhanced CT / MRI or PET-CT), doctors can intuitively track changes in tumor volume, blood perfusion characteristics and metabolic activity, and then quantitatively evaluate treatment response.

[0003] At present, dynamic image processing systems usually assist doctors in analysis in the following ways: the system allows doctors to freeze at any frame through manual operations (such as pausing playback, browsing frame by frame), and further measure the lesion size, shape or signal intensity of the current display frame through interactive tools (such as rulers, ROI delineation). However, such systems require doctors to manually screen key frames and perform measurements, which is inefficient in analysis, and the results are easily affected by differences in operator experience, attention fluctuations and fatigue, resulting in poor evaluation consistency. Summary of the invention

[0004] The main purpose of this application is to provide a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images, a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images and a storage medium, aiming to solve the technical problem of poor consistency of evaluation results in related technologies.

[0005] To achieve the above objectives, the present application provides a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images, the method comprising:

[0006] Receive a dynamic image, and based on a preset time step, capture at least two frames to be processed in the dynamic image;

[0007] Determine a first region of interest and a second region of interest in the picture to be processed, wherein the first region of interest is an abnormal tissue region, and the second region of interest is a non-abnormal tissue region;

[0008] Acquiring historical images based on the patient identification associated with the dynamic image;

[0009] grouping the historical pictures and the pictures to be processed according to the similarity between the second regions of interest in the historical pictures and the pictures to be processed;

[0010] A treatment effect evaluation result is generated according to the first region of interest of the historical picture and the picture to be processed in each picture group.

[0011] In the embodiment of the present application, the step of generating the treatment effect evaluation result according to the first region of interest of the historical picture and the picture to be processed in each picture group includes:

[0012] Based on the first regions of interest of the historical pictures and the pictures to be processed, extracting key evaluation indicators corresponding to the historical pictures and the pictures to be processed;

[0013] Constructing at least one evaluation key indicator sequence corresponding to the picture grouping according to the timestamps of the historical pictures and the pictures to be processed, wherein the evaluation key indicator sequence is sorted based on time;

[0014] The treatment effect evaluation result is generated based on the evaluation key indicator sequence.

[0015] In the embodiment of the present application, the key evaluation index includes the size of the lesion, and the step of extracting the key evaluation index corresponding to each of the historical screen and the screen to be processed based on the first region of interest of the historical screen and the screen to be processed includes:

[0016] determining a centroid position of the first region of interest;

[0017] According to the center of gravity position and the preset angle step, the size value corresponding to each evaluation angle direction is obtained as the key evaluation indicator.

[0018] In the embodiment of the present application, the evaluation key indicator sequence is a numerical sequence of the size values ​​corresponding to the same evaluation angle direction of each picture in the same picture group, which is sorted based on time.

[0019] In an embodiment of the present application, the key evaluation indicators include at least one of a lesion density value and a lesion size.

[0020] In the embodiment of the present application, the step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second region of interest in the pictures to be processed comprises:

[0021] Determine the cluster core corresponding to each of the picture groups;

[0022] Based on the clustering core, clustering is performed according to the similarity of the second regions of interest between pictures, and the historical pictures and the pictures to be processed are grouped according to the clustering result.

[0023] In the embodiment of the present application, the step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second region of interest in the pictures to be processed comprises:

[0024] Acquire n historical pictures closest to the current moment from the picture groups respectively;

[0025] Determine the similarity of the second region of interest between the picture to be processed and the n acquired historical pictures;

[0026] taking a weighted sum of the similarities between the second regions of interest as a grouping score of the picture to be processed relative to the picture group;

[0027] The to-be-processed pictures are classified into the picture groups according to the scores, and a grouping result including the historical pictures and the to-be-processed pictures is obtained.

[0028] In the embodiment of the present application, after the step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second regions of interest in the pictures to be processed, the method further includes:

[0029] The picture to be processed is associated with the grouping result and saved.

[0030] An embodiment of the present application also provides a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, wherein the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images comprises a memory, a processor, and a computer program stored on the memory, wherein the computer program, when executed by the processor, implements the steps of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as described above.

[0031] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as described above are implemented.

[0032] An embodiment of the present application discloses a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images. The method first receives a dynamic image, and based on a preset time step, captures at least two to-be-processed frames in the dynamic image, and further determines the abnormal tissue area and the non-abnormal tissue area in the to-be-processed frame, and then obtains a historical frame based on a patient identifier associated with the dynamic image, and groups the historical frame and the to-be-processed frame according to the similarity between the historical frame and the second region of interest in the to-be-processed frame, and finally generates a treatment effect evaluation result according to the first region of interest of the historical frame and the to-be-processed frame in each frame group, thereby achieving the purpose of automatically analyzing the cancer treatment effect evaluation result through artificial intelligence. Since the standards and rules are unified and fixed during the analysis process, the analysis results have better consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flowchart of an embodiment of a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images involved in an embodiment of the present application;

[0034] Figure 2 It is a structural schematic diagram of the region of interest analysis model involved in the embodiment of the present application;

[0035] Figure 3 This is a schematic diagram of the evaluation angle direction involved in the embodiment of the present application;

[0036] Figure 4 This is a schematic diagram of the structure of a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images in this application;

[0037] Figure 5 This is a schematic diagram of the modular structure of the cancer treatment evaluation system based on artificial intelligence analysis of dynamic images in this application.

[0038] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0039] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0040] Dynamic Medical Imaging is an imaging method that uses medical imaging technology to observe the changes in the internal structure, function or physiological process of the human body over time in real time or continuously. Compared with traditional static imaging technology (such as X-rays, conventional CT or MRI single scans), dynamic medical imaging can capture dynamic processes such as organ movement, hemodynamic changes, metabolic activity and drug distribution, thereby providing more comprehensive spatiotemporal information for clinical diagnosis and treatment. Especially in the field of cancer treatment, dynamic imaging technology is widely used in surgical planning, radiotherapy target positioning and staged evaluation of treatment effects. By collecting dynamic imaging data of the lesion area during different treatment cycles (such as dynamic enhanced CT / MRI or PET-CT), doctors can intuitively track changes in tumor volume, blood perfusion characteristics and metabolic activity, and then quantitatively evaluate treatment response.

[0041] At present, dynamic image processing systems usually assist doctors in analysis in the following ways: the system allows doctors to freeze at any frame through manual operations (such as pausing playback, browsing frame by frame), and further measure the lesion size, shape or signal intensity of the current display frame through interactive tools (such as rulers, ROI delineation). However, such systems require doctors to manually screen key frames and perform measurements, which is inefficient in analysis, and the results are easily affected by differences in operator experience, attention fluctuations and fatigue, resulting in poor evaluation consistency.

[0042] Based on the above defects of the existing system, the embodiment of the present application provides a cancer treatment evaluation method based on artificial intelligence analysis of dynamic images. The method first receives the dynamic image, and based on a preset time step, captures at least two to-be-processed images in the dynamic image, and further determines the abnormal tissue area and the non-abnormal tissue area in the to-be-processed image, and then obtains the historical image based on the patient identification associated with the dynamic image, and groups the historical image and the to-be-processed image according to the similarity of the second interest region in the historical image and the to-be-processed image, and finally generates the treatment effect evaluation result according to the first interest region of the historical image and the to-be-processed image in each image group, thereby achieving the purpose of automatically analyzing the cancer treatment effect evaluation result through artificial intelligence. Since the standards and rules are unified and fixed during the analysis process, the analysis results have better consistency.

[0043] For ease of understanding, the technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0044] Please refer to Figure 1 In an optional embodiment, the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images includes the following steps:

[0045] The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images provided in this embodiment can be executed in a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, or in an evaluation system composed of multiple independent devices that communicate with each other.

[0046] S10: receiving a dynamic image, and based on a preset time step, capturing at least two frames to be processed in the dynamic image;

[0047] The above dynamic images are based on dynamic medical images collected by corresponding medical detection equipment. The specific medical image content is not limited. For example, it can be based on continuous X-ray imaging, X-ray fluoroscopy, dynamic ultrasound, dynamic CT / MRI or nuclear medicine dynamic imaging, etc.

[0048] The execution subject of the method provided in this embodiment can receive dynamic images based on an interface directly connected to the medical detection equipment, or an interface connected to the corresponding dynamic image storage server.

[0049] After receiving the dynamic image, the subsequent analysis process can be executed immediately, or the subsequent analysis process can be triggered after receiving the corresponding analysis instruction. The analysis instruction can be actively triggered by the user based on the user interaction interface. For example, the user interaction interface is provided with an impact upload interface, through which the user uploads the dynamic image, and after the system receives the dynamic image, the analysis process can be triggered through the preset controls in the interface. Of course, in some variants, certain triggering rules can also be set in advance to automatically trigger the subsequent analysis process. For example, after receiving the dynamic image, it is added to the queue to be processed, and according to the first-in-first-out rule, the current processing capacity is used as the monitoring indicator, and the dynamic image is read from the queue to be processed for analysis in turn. This method can maximize the efficiency of dynamic image analysis when computing resources are limited.

[0050] After the processing process starts, at least two to-be-processed images can be captured from the dynamic image based on a preset time step. It can be understood that, for dynamic images, the images corresponding to the images are dynamic in time and space. That is, at different time points, there will be corresponding human body images collected from the same or different acquisition angles. Therefore, different time steps can be set based on the computing power of the system or according to the accuracy requirements of the analysis results. After the time step is selected, the images to be processed can be captured from the dynamic image at intervals of the preset time step.

[0051] It should be noted that, when a dynamic image contains multiple images corresponding to different viewing angles at the same time, the pictures corresponding to the different viewing angles at that time can be captured as the pictures to be processed.

[0052] Exemplarily, the preset time is not long and can be set to 10ms, 1s, 5s or 10s. If the selected preset time step is 1s, the corresponding picture frames at 61 moments such as the 0th second, the 1st second, ... the 59th second, the 60th second, etc. in the dynamic image with a duration of 1 minute can be intercepted as the pictures to be processed, thereby obtaining 61 pictures to be processed. Optionally, when the dynamic video includes pictures corresponding to multiple different perspectives, the number of pictures to be processed is n*m. Among them, n is the number of interception moments determined according to the preset time step, and m is the number of perspectives. For example, when the dynamic image has 2 different perspectives, 2 picture frames can be intercepted at each moment corresponding to the 0th second, the 1st second, ... the 59th second, and the 60th second, thereby obtaining 122 pictures to be processed.

[0053] Step S20: determining a first region of interest and a second region of interest in the image to be processed, wherein the first region of interest is an abnormal tissue region, and the second region of interest is a non-abnormal tissue region;

[0054] After obtaining the image to be processed, the abnormal tissue area, i.e., the lesion area, and the non-abnormal tissue area, i.e., the area outside the lesion, in each image to be processed can be determined by the region of interest analysis model. Hereinafter, the abnormal tissue area is described as the first region of interest, and the non-abnormal tissue area is described as the second region of interest.

[0055] It is understandable that in the image to be processed, only two areas are included, namely, the area outside the first area of ​​interest, which is defined as the second area of ​​interest. Therefore, the area of ​​interest analysis model only needs to identify the first area of ​​interest, and then after identifying the first area of ​​interest, automatically use the area outside the first area of ​​interest as the second perceptual area. This can effectively save computing power. Of course, in some schemes, if there is enough computing power, the first area of ​​interest and the second area of ​​interest can be identified separately. After obtaining the two results, the two results are fused based on confidence or other indicators to obtain the final first area of ​​interest and second area of ​​interest results. This can effectively improve the accuracy of the area recognition results.

[0056] For example, please refer to Figure 2 The region of interest analysis model can be an automated image analysis model based on deep learning, which can quickly and accurately locate the region of interest (ROI) from the input image. Its core modules include an input preprocessing module, a backbone network, and a region proposal module.

[0057] The input preprocessing module is used to standardize the original image (such as resizing and normalization), remove noise and enhance key features such as contrast stretching, etc., to ensure that images of different sources and qualities are suitable for subsequent analysis.

[0058] The backbone network can use convolutional neural networks, such as ResNet, EfficientNet, or visual Transformer, to extract low-level features of the image (such as edges and textures) layer by layer to high-level semantic features, such as organ contours, lesion morphology, etc.

[0059] The Region Proposal Module focuses on high-information areas, such as abnormal tissues in medical images, by calculating the saliency weights of each region in the feature map. Furthermore, it generates bounding boxes of potential ROIs by sliding a window on the feature map or using a region proposal network (RPN). It outputs a preliminary list of candidate regions, including location coordinates and confidence scores.

[0060] The ROI refinement module performs "boundary correction" on the candidate region, such as fine-tuning the bounding box through a regression network. It also performs multi-scale feature fusion to combine feature maps at different levels to improve the detection accuracy of small targets or blurred areas. Optionally, the contour details of the ROI can also be refined through a pixel-level segmentation network (such as U-Net).

[0061] Finally, the output module outputs the precise coordinates of the ROI (such as a rectangular box or mask) and gives a confidence score.

[0062] It should be noted that in the above-mentioned region of interest analysis model, the model gradually extracts image features through convolution or Transformer layers, and learns the implicit rules for distinguishing "important areas" from the background. For example, in medical images, tumor areas may show different texture, density or shape characteristics from normal tissues, and the model automatically captures these differences through labeled samples in the training data. Then, spatial attention or channel attention is used to dynamically allocate computing resources to key areas. For example, in lung CT images, the attention module will focus on high-density nodule areas first. After supervised learning through a large amount of labeled data (including ROI position labels), the loss function can simultaneously optimize positioning accuracy (such as the overlap IoU between the bounding box and the true box) and classification accuracy (such as distinguishing true and false positive areas). During the training process, the model gradually learns to balance sensitivity and specificity, reducing missed detections and false detections.

[0063] Step S30: acquiring historical images based on the patient identification associated with the dynamic image;

[0064] Step S40: grouping the historical pictures and the pictures to be processed according to the similarity between the second regions of interest in the historical pictures and the pictures to be processed;

[0065] In this embodiment, the patient identification can be obtained based on the attachment information of the received dynamic image. The patient identification is a unique identification code of the patient, which can be a serial code generated based on a preset coding rule, or can also be set to an ID number or other unique identification associated with the patient, which is not limited in this embodiment. After obtaining the patient identification, all historical screens associated with the patient identification can be obtained. It is understandable that when the system receives the dynamic image corresponding to the patient for the first time, the acquisition result of the historical screen is empty, then the subsequent grouping action can be directly performed based on the current screen to be processed to generate multiple screen groups. And the screen to be processed is used as a historical screen, and is saved in groups according to the grouping results.

[0066] Optionally, if the dynamic image corresponding to the patient is received for the first time, the subsequent step S50 is no longer performed after the grouping is completed. Indicators such as the time-varying law of the enhancement curve morphology and / or the blood flow rate can be extracted based on the dynamic image, and / or multi-dimensional feature fusion indicators such as metabolism-blood flow coupling changes can also be extracted, and then based on the extracted indicators and preset evaluation rules, the corresponding cancer initial state evaluation results are generated.

[0067] Optionally, when it is not the first time to receive a dynamic image, it can be further determined to which group each of the to-be-processed images belongs.

[0068] As an optional implementation scheme, the images to be processed can be grouped based on a clustering algorithm. For example, when a dynamic image is first received, each historical image corresponding to the first dynamic image is directly grouped separately, and then the second region of interest of each historical image corresponding to the first dynamic image is used as the clustering core to guide the subsequent grouping of the images to be processed. Alternatively, the first dynamic image can be corresponding to each historical image, and clustering is performed to obtain the grouping result during the first grouping. The fusion features of the second perceptual region of each historical image corresponding to the first dynamic image in the same group are used as the clustering core corresponding to the group to guide the subsequent grouping process of the images to be processed.

[0069] It can be understood that the clustering core is a vector expressing the second region of interest of a historical screen, or a fusion vector expressing the second region of interest of multiple historical screens. Therefore, in the distance process, the target vector expressing the second region of interest of each screen to be processed can be calculated. Then the similarity between each target vector and the distance core is calculated, and clustering is performed based on the similarity. And the grouping results of each screen to be processed are determined according to the clustering results. And after the grouping results are determined, they are saved in the corresponding group. That is, the screen to be processed is associated with the grouping result and saved.

[0070] In this embodiment, classification is performed through the second region of interest without referring to the first region of interest, so that the influence of the lesion area, i.e., the abnormal tissue area, on the corresponding classification result can be filtered out, thereby ensuring the accuracy of the sharing results in the subsequent analysis process.

[0071] In another optional implementation, the n historical images closest to the current moment may be obtained from the image group, that is, the n historical images closest to the acquisition event and the current moment, and then the similarity of the second region of interest between the image to be processed and the obtained n historical images is determined.

[0072] Optionally, in the process of determining the similarity between the picture to be processed and the n historical pictures, the present solution provides two optional methods. One is to first determine the n similarities between the n historical pictures one by one, then perform weighted summation on the n similarities, and use the weighted sum as the similarity result between the picture to be processed and the historical pictures in the group. Among them, the weight values ​​corresponding to the n historical pictures can increase from the front to the back with the acquisition time. Then, the weighted sum of the similarities is used as the grouping score of the picture to be processed relative to the picture group.

[0073] The second method is to first calculate the fusion vector based on the expression vectors of the second region of interest of the n historical pictures, and then use the similarity between the fusion vector and the expression vector of the second region of interest of the picture to be processed, that is, the target vector, as the similarity between it and the historical pictures in the group. And use the similarity as the grouping score of the picture to be processed relative to the picture group.

[0074] Then, the to-be-processed pictures are classified into the picture groups according to the scores, and a grouping result including the historical pictures and the to-be-processed pictures is obtained, that is, the to-be-processed pictures are divided into the groups with the highest grouping scores.

[0075] Step S50: generating a treatment effect evaluation result according to the first region of interest of the historical images and the images to be processed in each image group.

[0076] After the grouping is completed, each group may contain only historical pictures, or may contain both historical pictures and pictures to be processed. For groups that only contain historical pictures, it can be set not to process them to save computing power. Of course, this embodiment does not exclude the possibility of processing groups that only contain historical pictures at the same time. For ease of understanding and convenience of description, the historical pictures and / or pictures to be processed in each picture group after the grouping is completed are described as pictures within the group.

[0077] Based on the grouping results, the key evaluation indicators corresponding to the pictures in each group can be extracted based on the first region of interest of the pictures in the group corresponding to each group. Wherein, the key evaluation indicators include but are not limited to at least one of the lesion density value and the lesion size. For the lesion density, only one density value can be extracted from a picture in a group. Therefore, the lesion density values ​​corresponding to the pictures in each group in the same group of the lesion can be directly sorted in sequence according to the order from front to back in time based on the timestamps corresponding to the pictures in each group, thereby generating a lesion density sequence as a key evaluation indicator sequence. Therefore, each group can generate a lesion density sequence. Then, according to the effectiveness of each group, the lesion density sequence is fused. The density change of the patient's lesion area during the entire cancer treatment process is generated.

[0078] In the process of generating a key evaluation indicator sequence based on the lesion size, this implementation provides two optional implementation schemes. One is to define the lesion size directly based on the cross-sectional area or volume of the lesion area. Then, based on a similar method to the lesion density, a key evaluation indicator sequence is constructed.

[0079] Secondly, in the group screen, the center of gravity position of the first region of interest is first determined, and according to the center of gravity position and the preset angle step, the corresponding size value in each evaluation angle direction is obtained as the evaluation key indicator. Then, the evaluation key indicator sequence is constructed.

[0080] For example, please refer to Figure 3 After determining the center of gravity, the size values ​​corresponding to the 8 evaluation angle directions can be determined based on the center of gravity as the starting point and 45° as the angle compensation. That is, the size values ​​corresponding to the 8 evaluation angle directions of 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315° are determined. Then, based on the acquisition time sequence of the pictures in the group, the corresponding size values ​​in the same evaluation angle direction are sorted based on time to generate 8 numerical sequences. In this way, a group of pictures can obtain 8 numerical sequences. Furthermore, according to the obtained numerical sequence, the size changes corresponding to each evaluation angle direction in the group are obtained. And then based on the multiple groups of size changes corresponding to multiple groups of pictures, the size changes of the lesions during cancer treatment can be analyzed.

[0081] It is understandable that the angle step size can be a custom setting based on computing power or analysis requirements.

[0082] Furthermore, the system also constructs corresponding evaluation rules based on the treatment process change knowledge provided by cancer treatment experts in advance. After the change in lesion size and / or density is determined, the treatment effect evaluation result can be generated based on the evaluation rules and the determined change in lesion size and / or density. Alternatively, the change in lesion size and / or density is directly output as the treatment evaluation result.

[0083] In addition, after constructing the key evaluation indicator sequence, in the process of determining the change in lesion size and / or density according to the evaluation sequence, the change in lesion size and / or density can be generated based on the input key evaluation indicator sequence based on the pre-trained deep learning model. In the training process, the deep learning model can generate the key evaluation indicator sequence based on the sample data, annotate the corresponding results, and then train the model based on the annotated data.

[0084] In this embodiment, a dynamic image is first received, and based on a preset time step, at least two to-be-processed frames are captured in the dynamic image, and the abnormal tissue area and the non-abnormal tissue area in the to-be-processed frame are further determined, and then based on the patient identification associated with the dynamic image, a historical frame is obtained, and according to the similarity between the historical frame and the second region of interest in the to-be-processed frame, the historical frame and the to-be-processed frame are grouped, and finally, according to the first region of interest of the historical frame and the to-be-processed frame in each frame group, a treatment effect evaluation result is generated, thereby achieving the purpose of automatically analyzing the cancer treatment effect evaluation results through artificial intelligence, and since the standards and rules are unified and fixed during the analysis process, the analysis results have better consistency.

[0085] The present application provides a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, and the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images in the above-mentioned embodiment one.

[0086] Reference below Figure 4, which shows a schematic diagram of the structure of a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images suitable for implementing the embodiment of the present application. The cancer treatment evaluation device based on artificial intelligence analysis of dynamic images in the embodiment of the present application may include, but is not limited to, a server and a computer. Figure 4 The cancer treatment evaluation device based on artificial intelligence analysis of dynamic images shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0087] like Figure 4 As shown, the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM) 1004. In the random access memory 1004, various programs and data required for the operation of the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images are also stored. The processing device 1001, the read-only memory 1002 and the random access memory 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the input / output interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, etc.; an output device 1008 including, for example, a liquid crystal display (LCD, Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a cancer treatment evaluation device based on artificial intelligence analysis of dynamic images with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0088] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a read-only memory 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0089] The cancer treatment evaluation device based on artificial intelligence analysis of dynamic images provided by the present application adopts the cancer treatment evaluation device method based on artificial intelligence analysis of dynamic images in the above-mentioned embodiment to solve the technical problem of poor consistency of evaluation results. Compared with the relevant technology, the beneficial effects of the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images provided by the present application are the same as the beneficial effects of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images provided by the above-mentioned embodiment, and the other technical features of the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0090] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0092] Please refer to Figure 5 The present application provides a cancer treatment evaluation system based on artificial intelligence analysis of dynamic images. The cancer treatment evaluation system based on artificial intelligence analysis of dynamic images 100 includes:

[0093] The pre-processing module 110 receives a dynamic image and captures at least two frames to be processed in the dynamic image based on a preset time step;

[0094] A first analysis module 120 determines a first region of interest and a second region of interest in the image to be processed, wherein the first region of interest is an abnormal tissue region and the second region of interest is a non-abnormal tissue region;

[0095] The second analysis module 130 obtains historical images based on the patient identification associated with the dynamic image; and groups the historical images and the images to be processed according to the similarity between the historical images and the second regions of interest in the images to be processed; and generates treatment effect evaluation results according to the first regions of interest of the historical images and the images to be processed in each image group.

[0096] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images in the above-mentioned embodiment.

[0097] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM, Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM, CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0098] The computer-readable storage medium may be included in the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images; or it may exist independently without being assembled into the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images.

[0099] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, the cancer treatment evaluation device based on artificial intelligence analysis of dynamic images can improve the storage efficiency of the warehouse based on the above method.

[0100] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0101] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0102] The modules involved in the embodiments described in the present application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0103] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned cancer treatment evaluation method based on artificial intelligence analysis of dynamic images, which can solve the technical problem of poor consistency of evaluation results. Compared with the related art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images provided in the above-mentioned embodiment, and will not be repeated here.

[0104] An embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned cancer treatment evaluation method based on artificial intelligence analysis of dynamic images.

[0105] The computer program product provided in this application can solve the technical problem of poor consistency of evaluation results. Compared with the related art, the beneficial effects of the computer program product provided in the embodiment of this application are the same as the beneficial effects of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images provided in the above embodiment, which will not be repeated here.

[0106] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent processing scope of the present application.

[0107] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element. In the intervals given in this application, the boundary values ​​are all included unless explicitly defined.

[0108] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method.

[0109] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A cancer treatment evaluation method based on artificial intelligence analysis of dynamic images, characterized in that: The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images includes: Receive a dynamic image, and based on a preset time step, capture at least two frames to be processed in the dynamic image; Determine a first region of interest and a second region of interest in the picture to be processed, wherein the first region of interest is an abnormal tissue region, and the second region of interest is a non-abnormal tissue region; Acquiring historical images based on the patient identification associated with the dynamic image; grouping the historical pictures and the pictures to be processed according to the similarity between the second regions of interest in the historical pictures and the pictures to be processed; A treatment effect evaluation result is generated according to the first region of interest of the historical picture and the picture to be processed in each picture group.

2. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 1, characterized in that: The step of generating the treatment effect evaluation result according to the first region of interest of the historical picture and the picture to be processed in each picture group comprises: Based on the first regions of interest of the historical pictures and the pictures to be processed, extracting key evaluation indicators corresponding to the historical pictures and the pictures to be processed; Constructing at least one evaluation key indicator sequence corresponding to the picture grouping according to the timestamps of the historical pictures and the pictures to be processed, wherein the evaluation key indicator sequence is sorted based on time; The treatment effect evaluation result is generated based on the evaluation key indicator sequence.

3. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 2, characterized in that: The key evaluation index includes the size of the lesion, and the step of extracting the key evaluation index corresponding to each of the historical pictures and the picture to be processed based on the first region of interest of the historical pictures and the picture to be processed includes: determining a centroid position of the first region of interest; According to the center of gravity position and the preset angle step, the size value corresponding to each evaluation angle direction is obtained as the key evaluation indicator.

4. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 3, characterized in that: The evaluation key indicator sequence is a numerical sequence of size values ​​corresponding to the same evaluation angle direction of each picture in the same picture group, which is sorted based on time.

5. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 2, characterized in that: The key evaluation index includes at least one of a lesion density value and a lesion size.

6. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 1, characterized in that: The step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second regions of interest in the pictures to be processed comprises: Determine the cluster core corresponding to each of the picture groups; Based on the clustering core, clustering is performed according to the similarity of the second regions of interest between pictures, and the historical pictures and the pictures to be processed are grouped according to the clustering result.

7. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 1, characterized in that: The step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second regions of interest in the pictures to be processed comprises: Acquire n historical pictures closest to the current moment from the picture groups respectively; Determine the similarity of the second region of interest between the picture to be processed and the n acquired historical pictures; taking a weighted sum of the similarities between the second regions of interest as a grouping score of the picture to be processed relative to the picture group; The to-be-processed pictures are classified into the picture groups according to the scores, and a grouping result including the historical pictures and the to-be-processed pictures is obtained.

8. The cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as claimed in claim 7, characterized in that: After the step of grouping the historical pictures and the pictures to be processed according to the similarity between the historical pictures and the second regions of interest in the pictures to be processed, the method further includes: The picture to be processed is associated with the grouping result and saved.

9. A cancer treatment evaluation device based on artificial intelligence analysis of dynamic images, characterized in that: The cancer treatment evaluation device based on artificial intelligence analysis of dynamic images includes a memory, a processor, and a computer program stored on the memory. When the computer program is executed by the processor, the steps of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as described in any one of claims 1 to 8 are implemented.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the cancer treatment evaluation method based on artificial intelligence analysis of dynamic images as described in any one of claims 1 to 8 are implemented.

Images