Cancer pain treatment response evaluation method based on image analysis

Abstract

The invention relates to the field of cancer pain analysis of human body images, in particular to a cancer pain treatment response evaluation method based on image analysis. The method comprises the steps of collecting continuous human body images of a patient; firstly, angular points in each image are collected; obtaining an angular point aggregation coefficient according to the distance characteristics between the angular points; uniformly dividing the image to obtain human body image layers; in the layering process, key angular points are obtained according to the angular point aggregation coefficient difference of the angular points; obtaining an activity coefficient according to similar features between the angular point aggregation coefficient difference and the key angular points; obtaining a response degree according to the activity coefficient; and performing response evaluation on cancer pain treatment according to the response degree of the patient. According to the invention, a non-invasive, real-time and quantitative cancer pain treatment dynamic evaluation method can be provided for the patient by dynamically tracking the activity characteristics of the patient.

Description

Technical Field

[0001] The present invention relates to the field of cancer pain analysis of human body images, and in particular to a method for evaluating cancer pain treatment response based on image analysis. Background Art

[0002] Currently, assessment of cancer pain response relies primarily on subjective patient reports and physicians' clinical judgment, including pain rating scales such as the visual analog scale and clinical guidelines. However, these methods have limitations, such as high subjectivity and a lack of objective quantitative indicators. In recent years, with the development of image processing technologies, cancer pain response assessment using image analysis has gained increasing attention. Image analysis can reduce interference from human factors and improve the reliability and repeatability of assessments, thereby providing patients with more objective and accurate assessment results and enabling more accurate judgment of treatment efficacy.

[0003] The evaluation of the effectiveness of cancer pain treatment is usually periodic, and real-time monitoring and dynamic evaluation are crucial for the treatment of cancer pain. However, due to the differences in cancer pain treatment goals and individuals among patients, it is a technical challenge to achieve non-invasive, real-time dynamic evaluation of the response to cancer pain treatment based on individual differences among patients. Summary of the Invention

[0004] In order to solve the technical problem that it is difficult to conduct non-invasive, real-time dynamic assessment of cancer pain treatment based on individual differences among patients due to differences in cancer pain treatment goals, the present invention aims to provide a method for assessing cancer pain treatment response based on image analysis. The technical solution adopted is as follows: a method for assessing cancer pain treatment response based on image analysis, the method comprising: collecting human body images of patients at all sampling moments; performing corner point detection on all the human body images to obtain all corner points of each human body image; selecting a corner point in any human body image as a reference corner point; and determining the response of the patient to the treatment based on the distance between the reference corner point and a preset number of other nearest corner points. features, obtain the corner clustering coefficient of the reference corner points; evenly divide each of the human body images to obtain human body image layers; in each of the human body image layers, obtain the key corner points in the human body image according to the corner clustering coefficient of each corner point; obtain the activity coefficient of the patient between two adjacent sampling moments according to the difference in corner clustering coefficients of the key corner points in the human body images at two adjacent sampling moments and the similar features between the key corner points in the two human body images; obtain the patient's response degree according to the change characteristics of the patient's activity coefficient corresponding to the human body images at each two adjacent sampling moments; and evaluate the response of cancer pain treatment according to the patient's response degree.

[0005] Furthermore, the method for obtaining the corner point clustering coefficient includes: in the human body image, calculating the distance between the reference corner point and each other corner point as the first distance; selecting a preset number of other corner points with the closest first distance as adjacent corner points; calculating the mean of the first distance between the adjacent corner points and the reference corner point, and performing a negative correlation normalization operation to obtain the corner point clustering coefficient of the reference corner point.

[0006] Furthermore, each of the human body images is evenly divided to obtain human body image layers, including: evenly dividing each human body image from top to bottom into a preset first number of parts, and using each of the divided human body images as the human body image layer.

[0007] Furthermore, the method for obtaining key corner points includes: in each human body image layer, taking the corner point with the largest corner point clustering coefficient as the key corner point in the human body image layer; traversing all human body image layers in the human body image to obtain all key corner points in the human body image.

[0008] Furthermore, the method for obtaining the patient's activity coefficient includes: calculating the sum of the corner point clustering coefficients of all key corner points in the human body image at each sampling moment as the patient state coefficient of the human body image at each sampling moment; calculating the DTW value between all key corner points in the human body images at two adjacent sampling moments as the activity state similarity coefficient of the two adjacent sampling moments; taking the sum of the patient state coefficient of the latter human body image in the two adjacent sampling moments and the activity state coefficient of the two adjacent sampling moments as the first sum; taking the ratio of the first sum to the patient state coefficient of the previous human body image in the two adjacent sampling moments as the patient's activity coefficient between the two adjacent sampling moments.

[0009] Furthermore, the method for obtaining the patient's response level includes: taking the latter sampling moment of each two adjacent sampling moments as the sampling moment where the activity coefficient is located; sorting the activity coefficients between each two adjacent sampling moments of the patient in time sequence, and clustering all activity coefficients into two clusters, taking the activity coefficient mean result of the two clusters with a larger mean value as the active cluster, and taking the activity coefficient mean result of the two clusters as the inactive cluster; sorting each activity coefficient in the active cluster and the inactive cluster in time sequence; and obtaining the response level according to a response level calculation formula, which is as follows: Where, Indicates the patient's response level; Indicates the number of activity coefficients contained in the active cluster; Indicates the number of active clusters activity coefficient; Indicates the maximum activity coefficient in the active cluster; Indicates the number of active clusters The sampling moment of the activity coefficient; Indicates the number of active clusters The sampling moment of the activity coefficient; represents the standard deviation of the activity coefficient in the active cluster.

[0010] Furthermore, based on the patient's response level, a response evaluation of cancer pain treatment is performed, including: when the patient's response level is greater than a preset first threshold, it is considered that the cancer pain treatment effect is obvious; when the patient's response level is less than a preset second threshold, it is considered that the cancer pain treatment effect is poor.

[0011] A cancer pain treatment response assessment system based on image analysis, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned cancer pain treatment response assessment method based on image analysis when executing the computer program.

[0012] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for evaluating cancer pain treatment response based on image analysis.

[0013] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the above-mentioned method for evaluating cancer pain treatment response based on image analysis are implemented.

[0014] The present invention has the following beneficial effects: the present invention collects body images of patients at all sampling moments, records the complete temporal changes of patients' movements through continuous body images, and captures subtle movements; since the corner points of body images usually correspond to special parts of the body of cancer patients, such as joints, hands and feet, the activity characteristics of cancer patients can be analyzed by analyzing the position distribution of body images of cancer patients' joints at continuous sampling moments, so all corner points of each body image are obtained; there are a certain number of useless corner points among the obtained corner points, which cannot reflect the activity characteristics of cancer patients, so all the obtained corner points need to be screened to obtain key corner points; when the joints of cancer patients are in a state of not stretching, the key corner points are obtained. When the body is in a state of being stretched or the limbs have a tendency to approach each other, the corner points of the joints in the corresponding human body image show a certain degree of clustering, so the corner point clustering coefficient of the corner points is analyzed; if the first few corner points with the largest corner point clustering coefficient are used as key corner points, it may lead to the emergence of local optimal solutions, so the human body image is evenly divided to avoid the emergence of local optimal solutions; the position changes of the key corner points in the human body image at each two adjacent sampling moments can reflect the patient's mobility, and the patient's cancer pain treatment situation can be evaluated by the mobility; so the patient's activity coefficient between two adjacent sampling moments is analyzed; according to the patient's response level, the response to cancer pain treatment is evaluated. The present invention can provide patients with a non-invasive, real-time and quantitative cancer pain treatment dynamic evaluation method by dynamically tracking the patient's activity characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 A flowchart of a method for evaluating cancer pain treatment response based on image analysis is provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0017] To further illustrate the technical means and effectiveness of the present invention in achieving its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of an image analysis-based cancer pain treatment response assessment method proposed by the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The following describes in detail a method for evaluating cancer pain treatment response based on image analysis provided by the present invention with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a method for evaluating cancer pain treatment response based on image analysis provided by an embodiment of the present invention, the method comprising: Step S1: collecting body images of the patient at all sampling moments.

[0021] The embodiments of the present invention are primarily used in scenarios where the effectiveness of cancer pain treatment is evaluated. Because cancer pain patients experience pain during activities and have a limited range of motion, to avoid subjective errors in observation by relevant personnel, the effectiveness of cancer pain treatment is assessed in real time based on the patient's limb movements. Therefore, in this embodiment, body images of the patient are collected at all sampling moments. This continuous body image captures the complete temporal changes in the patient's movements, capturing subtle movements.

[0022] In one embodiment of the present invention, initial human body images are collected at each sampling time within 24 hours after cancer pain treatment from a cancer pain patient. These initial human body images are grayscaled and denoised. Grayscale and denoising methods are well known to those skilled in the art and are not limited or detailed here. In one embodiment of the present invention, the sampling time is set to 1 second, meaning that human body images are collected at consecutive sampling times. It should be noted that the sampling time can be set arbitrarily and is not limited here.

[0023] Step S2: perform corner point detection on all human body images to obtain all corner points of each human body image; select any corner point in any human body image as a reference corner point; obtain the corner point clustering coefficient of the reference corner point based on the distance characteristics between the reference corner point and a preset number of other nearest corner points; evenly divide each human body image to obtain human body image layers; in each human body image layer, obtain the key corner points in the human body image based on the corner point clustering coefficient of each corner point; obtain the activity coefficient of the patient between two adjacent sampling moments based on the difference in corner point clustering coefficients of the key corner points in the human body images at two adjacent sampling moments and the similarity characteristics between the key corner points in the two human body images; obtain the patient's response degree based on the change characteristics of the patient's activity coefficient corresponding to the human body images at each two adjacent sampling moments.

[0024] In actual situations, the corner points of a human body image usually correspond to special parts of a cancer patient's body, such as joints, hands and feet. By analyzing the position distribution of the human body images of cancer patients' joints at consecutive sampling moments, the activity characteristics of the cancer patients can be analyzed. Therefore, in an embodiment of the present invention, corner point detection is performed on all human body images to obtain all corner points of each human body image.

[0025] In one embodiment of the present invention, the ORB corner detection algorithm is used to detect corners in all human images. It should be noted that the Harris corner detection algorithm can also be used, which is not specified here. The ORB corner detection algorithm is a well-known technique for those skilled in the art and is not limited or elaborated on here.

[0026] However, in practice, a certain number of acquired corner points may be useless and fail to reflect the activity characteristics of cancer patients. Therefore, all acquired corner points need to be screened. When a cancer patient's joints are not extended, or their limbs tend to be close together, the corner points at the joints in the corresponding human body image will show a certain degree of clustering. If there are corner points that show clustering, these corner points are considered to be key corner points. Therefore, in this embodiment of the present invention, the corner point clustering coefficient of the reference corner point is calculated based on the distance characteristics between the reference corner point and a preset number of other closest corner points.

[0027] Preferably, in one embodiment of the present invention, a method for obtaining a corner clustering coefficient includes: calculating the distance between a reference corner point and each other corner point in a human body image as a first distance; selecting a preset number of other corner points with the closest first distance as adjacent corner points; calculating the mean of the first distances between the adjacent corner points and the reference corner point, and performing a negative correlation normalization operation to obtain a corner clustering coefficient for the reference corner point. In one embodiment of the present invention, the corner clustering coefficient calculation formula is as follows: Where, The corner clustering coefficient representing the reference corner point; Indicates the preset number of other corner points that are closest to the reference corner point. In one embodiment of the present invention, the preset number is set to 5. The preset number can be set arbitrarily and is not limited here; The closest point to the reference corner The distances between the other corner points and the reference corner point; Represents an exponential function with a natural constant as its base.

[0028] In the corner point clustering coefficient calculation formula, the smaller the mean distance between other corner points and the reference corner point, the more clustered the corner points in a smaller range around the reference corner point are, and the larger the corner point clustering coefficient of the reference corner point is.

[0029] Since the larger the corner clustering coefficient is, the more obvious the clustering of the corner point with the surrounding corner points is, but near a corner with a large corner clustering coefficient, there is more likely to be another corner with a relatively large corner clustering coefficient. If the first few corner points with the largest corner clustering coefficient are used as key corner points, it may lead to the emergence of a local optimal solution, that is, key corner points are more likely to appear only in certain joints of the patient with obvious flexion and extension. However, the evaluation of the patient's cancer pain treatment response needs to be judged based on the patient's overall range of motion, so it is necessary to avoid the emergence of a local optimal solution. After stratification, selecting the corner with the largest clustering coefficient as the key corner point for each layer can enable the local area after the overall division of the patient to obtain the local maximum value, rather than falling into a local optimal solution in only a certain local area. In order to better understand the local features in human body images, in an embodiment of the present invention, each human body image is evenly divided to obtain human body image stratification.

[0030] Preferably, in one embodiment of the present invention, each human body image is evenly divided from top to bottom into a preset first number of portions, and each portion of the divided human body image is used as a human body image layer. It should be noted that the preset first number is set to 100, and the preset first number can be set arbitrarily and is not limited here.

[0031] Research is conducted in each human image layer to obtain key corner points. Preferably, in one embodiment of the present invention, the method for obtaining key corner points includes: selecting, in each human image layer, the corner point with the largest corner point clustering coefficient as the key corner point in that human image layer; and traversing all human image layers in the human image to obtain all key corner points in the human image.

[0032] The changes in the positions of key corner points in human body images between two adjacent sampling moments can reflect a patient's mobility, which can be used to assess the patient's cancer pain treatment. A narrow range of motion indicates a significant impact on the patient; conversely, a wide range indicates a minimal impact. Active participation in daily activities indicates a positive treatment outcome. Therefore, in this embodiment of the present invention, the patient's activity coefficient between two adjacent sampling moments is derived based on the difference in the corner clustering coefficients of key corner points within human body images at two adjacent sampling moments, as well as the similarity between the key corner points within the two images.

[0033] Preferably, in one embodiment of the present invention, the method for obtaining the patient's activity coefficient includes: calculating the sum of the corner point clustering coefficients of all key corner points in the human body image at each sampling moment as the patient state coefficient of the human body image at each sampling moment.

[0034] The DTW values ​​between all key corner points in the human body images at two adjacent sampling moments are calculated as the activity state similarity coefficients at the two adjacent sampling moments.

[0035] The sum of the patient state coefficient of the latter human body image at two adjacent sampling moments and the activity state coefficient at two adjacent sampling moments is taken as the first sum value; the ratio of the first sum value to the patient state coefficient of the former human body image at two adjacent sampling moments is taken as the activity coefficient of the patient between the two adjacent sampling moments.

[0036] In one embodiment of the present invention, the activity coefficient calculation formula is as follows: Where, represents the patient's activity coefficient; Indicates the number of key corner points of the latter human body image in two adjacent sampling moments; Indicates the first human body image in the second human body image between two adjacent sampling moments. Corner point clustering coefficient of key corner points; Indicates the number of key corner points of the previous human body image in the human body images at two adjacent sampling moments; Indicates the first human body image in the previous human body image between two adjacent sampling moments. Corner point clustering coefficient of key corner points; The patient state coefficient representing the latter human body image among the human body images at two adjacent sampling moments; Represents the patient status coefficient in the previous human body image between two adjacent sampling moments; Represents the activity state coefficient of two adjacent sampling moments.

[0037] In the activity coefficient calculation formula, the corner point clustering coefficient of the key corner points in the human body image of the next two adjacent sampling moments and The sum of the corner clustering coefficients of the key corner points in the previous human body image at the two adjacent sampling moments The changes between can reflect the changes in the patient's activity position between two adjacent sampling moments; if Relatively large, and When the value is relatively small, it means that the patient tends to change his posture at two adjacent sampling moments, reflecting that the patient tends to have a more ideal mobility, so the corresponding activity coefficient is larger; the human body image layers of the human body images at two adjacent sampling moments are consistent. If the patient has limb movements, the key corner points belonging to the same layer will change in the two images. The more layers that interact, the more ideal the patient's mobility is, reflecting a positive response to cancer pain treatment. However, even if the key corner points are in the same layer, they may not be aligned in the same layer, because there are multiple rows of pixels in a layer horizontally, and the corner points in the same layer may not be in the same row. Therefore, it is not advisable to use the Pearson correlation coefficient. Therefore, the present invention adopts dynamic time planning, that is, The value makes it possible to calculate the minimum alignment path even if the human body image layers are not aligned, thereby measuring the similarity of the key corner points of the human body images at two adjacent sampling moments. The smaller the value, the more similar it is and the smaller the activity coefficient is, and vice versa.

[0038] Because cancer pain treatment is an ongoing process, it's necessary to evaluate a patient's response to cancer pain treatment using the activity coefficients obtained at each adjacent sampling moment. Time-series changes in a patient's activity coefficient reflect changes in their enthusiasm for cancer pain treatment. Large and frequent changes in the activity coefficient indicate a significant increase in the patient's enthusiasm for cancer pain treatment, while a decrease suggests a decrease in their enthusiasm. Therefore, it's necessary to define the patient's response based on the time-series changes in the activity coefficient.

[0039] Preferably, in one embodiment of the present invention, the patient's response degree is obtained based on the patient's activity coefficient corresponding to the human body image at all sampling moments. The method includes: first, sorting the patient's activity coefficients between each two adjacent sampling moments in time sequence, and clustering all activity coefficients into two clusters, and taking the activity coefficient mean result of the two clusters with a larger one as the active cluster, and taking the activity coefficient mean result of the two clusters with a smaller one as the inactive cluster. The method includes: since the patient's activity state can be divided into two categories, active state and inactive state, the patient's activity coefficient is clustered to complete the classification, and the data object of the initial cluster is randomly extracted from the obtained activity coefficients. The reason for not using sequential extraction according to continuous time series is that random extraction can avoid the situation where the difference between classes after data clustering is not obvious due to the single state of the patient. The clustering algorithm used in this invention is Clustering, and specify the number of classification clusters for the algorithm as , after clustering is completed, the difference in activity coefficients in each cluster is small, and The activity coefficients between clusters are quite different. It should be noted that the activity coefficients in each cluster should be sorted in time sequence. The mean of the activity coefficient in each cluster is calculated respectively. The cluster corresponding to the larger mean result is called the active cluster, and the other cluster corresponding to the smaller mean result is called the inactive cluster.

[0040] The response degree is obtained according to the response degree calculation formula. The response degree calculation formula is as follows: Where, Indicates the patient's response level; Indicates the number of activity coefficients contained in the active cluster; Indicates the number of active clusters activity coefficient; Indicates the maximum activity coefficient in the active cluster; Indicates the number of active clusters The sampling moment of the activity coefficient; Indicates the number of active clusters The sampling moment of the activity coefficient; represents the standard deviation of the activity coefficient in the active cluster.

[0041] In the response degree calculation formula, the number of activity coefficients contained in the active cluster The more the number of activity coefficients, the more obvious the tendency of the patient to be in a state of obvious activity, and the greater the patient's response at this time; the larger the mean value of the activity coefficient in the active cluster, and the smaller the difference between it and the maximum value of the activity coefficient, the more obvious the tendency of the patient to be in a state of obvious activity, and the greater the patient's response at this time; the smaller the difference between the corresponding sampling moments of two adjacent activity coefficients in the active cluster, the greater the continuity of the patient's activity coefficient, and the greater the patient's response at this time; for the inactive cluster, more attention is paid to the stability of the patient. The better the stability of the patient in the inactive state, the more effective the cancer pain is, and the basal pain is at a low level, which is in line with the goal of painless rest in cancer pain treatment. The smaller the value of The larger the value.

[0042] Step S3: Evaluate the response to cancer pain treatment based on the patient's response level.

[0043] Preferably, in one embodiment of the present invention, if the response level is improved to a high level, for example, the response level improvement compared to the previous treatment exceeds a preset first threshold, it indicates that the cancer pain treatment effect is ideal and the patient's symptoms are significantly relieved. A subsequent follow-up plan can be designed to closely monitor the patient's pain changes to ensure that there is no rebound in the treatment.

[0044] If the response level is low, for example, the response level improvement compared to the previous treatment is less than the preset second threshold, it indicates that the treatment effect is not ideal and the effectiveness of the current treatment plan should be reviewed. The treatment method may need to be adjusted or more aggressive treatment methods may be adopted. It is also possible to consider combining imaging data and the patient's pathological information to assess whether further examination or diagnosis is needed. For example, factors such as tumor resistance and disease progression should be considered.

[0045] In one embodiment of the present invention, the preset first threshold is set to 20%, and the preset second threshold is set to 10%. It should be noted that the preset first threshold and the preset second threshold can be set arbitrarily and are not limited here.

[0046] At this point, the evaluation of cancer pain treatment response is completed.

[0047] In summary, human body images of patients at all sampling moments are collected; corner point detection is performed on all human body images to obtain all corner points of each human body image; a corner point is randomly selected in any human body image as a reference corner point; the corner point clustering coefficient of the reference corner point is obtained based on the distance characteristics between the reference corner point and a preset number of other nearest corner points; each human body image is evenly divided to obtain human body image layers; in each human body image layer, the key corner points in the human body image are obtained based on the corner point clustering coefficient of each corner point; the activity coefficient of the patient between two adjacent sampling moments is obtained based on the difference in corner point clustering coefficients of the key corner points in the human body images at two adjacent sampling moments and the similar features between the key corner points in the two human body images; the patient's response level is obtained based on the patient's activity coefficient corresponding to the human body images at all sampling moments; and the response evaluation of cancer pain treatment is performed based on the patient's response level.

[0048] The second purpose of one embodiment of the present invention is to provide a cancer pain treatment response assessment system based on image analysis, the system comprising a memory, a processor and a computer program, wherein the memory is used to store the corresponding computer program, the processor is used to run the corresponding computer program, and when the computer program runs in the processor, it can implement the method described in steps S1-S3.

[0049] The third object of an embodiment of the present invention is to provide a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method described in steps S1-S3 is implemented when the processor executes the computer program.

[0050] A fourth object of an embodiment of the present invention is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in steps S1-S3 is implemented.

[0051] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0052] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for evaluating cancer pain treatment response based on image analysis, characterized in that: The method includes: collecting human body images of the patient at all sampling moments; performing corner point detection on all the human body images to obtain all corner points of each human body image; selecting any corner point in any human body image as a reference corner point; obtaining a corner point clustering coefficient of the reference corner point based on the distance characteristics between the reference corner point and a preset number of other corner points closest to the reference corner point; uniformly dividing each human body image to obtain human body image layers; in each human body image layer, obtaining key corner points in the human body image based on the corner point clustering coefficient of each corner point; obtaining the activity coefficient of the patient between two adjacent sampling moments based on the difference in corner point clustering coefficients of the key corner points in the human body images at two adjacent sampling moments and the similarity characteristics between the key corner points in the two human body images; obtaining the patient's response level based on the change characteristics of the patient's activity coefficient corresponding to the human body images at each two adjacent sampling moments; and performing a response evaluation on cancer pain treatment based on the patient's response level.

2. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, characterized in that: The method for obtaining the corner point clustering coefficient includes: in the human body image, calculating the distance between the reference corner point and each other corner point as the first distance; selecting a preset number of other corner points with the closest first distance as adjacent corner points; calculating the mean of the first distances between the adjacent corner points and the reference corner point, and performing a negative correlation normalization operation to obtain the corner point clustering coefficient of the reference corner point.

3. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, characterized in that: Each of the human body images is evenly divided to obtain human body image layers, including: evenly dividing each human body image from top to bottom into a preset first number of parts, and using each of the divided human body images as the human body image layer.

4. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, wherein: The method for obtaining key corner points includes: in each human body image layer, taking the corner point with the largest corner point clustering coefficient as the key corner point in the human body image layer; traversing all human body image layers in the human body image to obtain all key corner points in the human body image.

5. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, characterized in that: The method for obtaining the patient's activity coefficient includes: calculating the sum of the corner point clustering coefficients of all key corner points in the human body image at each sampling moment as the patient state coefficient of the human body image at each sampling moment; calculating the DTW value between all key corner points in the human body images at two adjacent sampling moments as the activity state similarity coefficient of the two adjacent sampling moments; taking the sum of the patient state coefficient of the latter human body image in the two adjacent sampling moments and the activity state coefficient of the two adjacent sampling moments as a first sum value; and taking the ratio of the first sum value to the patient state coefficient of the former human body image in the two adjacent sampling moments as the patient's activity coefficient between the two adjacent sampling moments.

6. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, characterized in that: The method for obtaining the patient's response level includes: using the later sampling moment between each two adjacent sampling moments as the sampling moment where the activity coefficient is located; sorting the activity coefficients between each two adjacent sampling moments of the patient according to time sequence, and clustering all activity coefficients into two clusters, using the activity coefficient mean result of the two clusters with a larger mean value as the active cluster, and using the activity coefficient mean result of the two clusters with a smaller mean value as the inactive cluster; sorting each activity coefficient in the active cluster and the inactive cluster according to time sequence; and obtaining the response level according to a response level calculation formula, wherein the response level calculation formula is as follows: Where, Indicates the patient's response level; Indicates the number of activity coefficients contained in the active cluster; Indicates the number of active clusters activity coefficient; Indicates the maximum activity coefficient in the active cluster; Indicates the number of active clusters The sampling moment of the activity coefficient; Indicates the number of active clusters The sampling moment of the activity coefficient; represents the standard deviation of the activity coefficient in the active cluster.

7. The method for evaluating cancer pain treatment response based on image analysis according to claim 1, characterized in that: Based on the patient's response level, a response evaluation of cancer pain treatment is performed, including: when the patient's response level is greater than a preset first threshold, it is considered that the cancer pain treatment effect is obvious; when the patient's response level is less than a preset second threshold, it is considered that the cancer pain treatment effect is poor.

8. A cancer pain treatment response assessment system based on image analysis, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for evaluating cancer pain treatment response based on image analysis as described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for evaluating cancer pain treatment response based on image analysis as claimed in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method for evaluating cancer pain treatment response based on image analysis as described in any one of claims 1 to 7 are implemented.

Images