Information processing method and device, computer device and computer readable storage medium

Abstract

The application relates to the technical field of data processing, is suitable for the medical and health field, and provides an information processing method and device, computer equipment and a computer readable storage medium, the method comprising the following steps: acquiring a first image sequence; if a target living body exists in the first image sequence, extracting a second image sequence in the first image sequence based on the similarity of each group of adjacent first images in the first image sequence; acquiring a skeletal feature vector of each second image in the second image sequence based on a predetermined skeletal recognition algorithm, so as to obtain a skeletal feature sequence; determining the motion information of each skeletal point of the target living body based on the skeletal feature sequence; and determining the activity amount of the target living body based on the motion information of each skeletal point of the target living body, wherein the activity amount is used to calculate the motion characteristic information of the target living body. The technical scheme can flexibly calculate the activity amount of a user by combining the skeletal point motion in the position and posture change of the skeletal points of the user.

Description

[0001] The present application relates to the technical field of data processing, and is suitable for the field of medical health, in particular to an information processing method and device, computer equipment and computer readable storage medium.

[0002] With the rapid development of science and technology, digital medical treatment gradually integrates into people's life, and a series of health services for users are derived. Among them, the portable devices such as mobile phones and wearable devices are often provided with health services with calorie calculation function for users, and other devices providing health services also often need to use the data of user's calorie when processing related information. Therefore, the accuracy of calorie calculation is particularly important.

[0003] An important parameter in calorie calculation is the activity amount of the user. In related technologies, the overall motion trajectory of the user is generally collected by wearable devices such as smart bracelets and smart wristbands, the overall motion trajectory of the user is converted into activity amount, and the value of calorie is obtained by substituting the activity amount into a preset formula.

[0004] However, this calculation method is too simple, and the overall motion trajectory of the user cannot fully reflect the activity amount of the user.

[0005] For example, the user exercises in place by slow movement such as yoga, in which case the user does not move but has a large activity amount. Therefore, the way of simply converting the overall motion trajectory of the user into activity amount affects the accuracy of the final calorie calculation.

[0006] Therefore, how to accurately and effectively calculate the activity amount of the user has become a technical problem to be solved at present.

[0007] The embodiments of the present application provide an information processing method and device, computer equipment and computer readable storage medium, which aims to solve the technical problem that the way of converting the overall motion trajectory of the user into activity amount in related technologies cannot meet the accuracy requirement of calorie calculation.

[0008] ​​​In a first aspect, an embodiment of the present application provides an information processing method, comprising: acquiring a first image sequence, the first image sequence comprising continuous frame images in a target video; if there is a target living body in the first image sequence, extracting a second image sequence in the first image sequence based on the similarity of each group of adjacent first images in the first image sequence; acquiring a skeletal feature vector of each second image in the second image sequence based on a predetermined skeletal recognition algorithm, to obtain a skeletal feature sequence; determining motion information of each skeletal point of the target living body based on the skeletal feature sequence, wherein the motion information comprises a current position of the skeletal point in each skeletal feature vector and a confidence of the current position, the confidence being calculated by an OKS algorithm and used to reflect the similarity between the current position of the skeletal point and a preset position; and determining an activity amount of the target living body based on the motion information of each skeletal point of the target living body, the activity amount being used to calculate motion feature information of the target living body.

[0009] In a second aspect, an embodiment of the present application provides an information processing apparatus, comprising: a first image sequence acquisition unit configured to acquire a first image sequence, the first image sequence comprising continuous frame images in a target video; a second image sequence acquisition unit configured to, if there is a target living body in the first image sequence, extract a second image sequence in the first image sequence based on the similarity of each group of adjacent first images in the first image sequence; a skeletal feature sequence acquisition unit configured to acquire a skeletal feature vector of each second image in the second image sequence based on a predetermined skeletal recognition algorithm, to obtain a skeletal feature sequence; a skeletal point motion information determination unit configured to determine motion information of each skeletal point of the target living body based on the skeletal feature sequence, wherein the motion information comprises a current position of the skeletal point in each skeletal feature vector and a confidence of the current position, the confidence being calculated by an OKS algorithm and used to reflect the similarity between the current position of the skeletal point and a preset position; and an activity amount determination unit configured to determine an activity amount of the target living body based on the motion information of each skeletal point of the target living body, the activity amount being used to calculate motion feature information of the target living body.

[0010] In a third aspect, an embodiment of the present application provides a computer device, comprising: 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 configured to execute the method in the first aspect.

[0011] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium storing computer executable instructions, the computer executable instructions being used to execute the method in the first aspect.

[0012] The above technical solution, in order to solve the problem that the way of converting the overall motion trajectory of the user to the activity amount in the related art cannot meet the accuracy requirement of calorie calculation, first, a first image sequence is obtained, the first image sequence includes consecutive frame images in a target video. The first image sequence is a video frame of a video. Specifically, a video can be collected by any computer device capable of video collection such as a mobile phone, a wearable device, a monitoring device, etc. The first image sequence composed of video frames in the video is used as the basis for activity amount calculation.

[0013] Then, if the first image sequence has a target living body, a second image sequence is extracted from the first image sequence based on the similarity of each group of adjacent first images in the first image sequence. The target living body refers to a predetermined object that needs to be calculated for activity amount, such as a user of a wearable device. The similarity of adjacent first images reflects the activity of the target living body, so the second image sequence can be composed of second images in which the body posture of the target living body changes based on the similarity of each group of adjacent first images in the first image sequence. That is, the second image sequence is an image in which the body posture of the target living body changes, reflecting the activity of the target living body.

[0014] Then, based on a predetermined bone recognition algorithm, the bone feature vector of each second image in the second image sequence is obtained to obtain a bone feature sequence. The change of the bone feature vector of each second image in the bone feature sequence reflects the posture change of the target living body reaching the activity amount statistical standard. Further, based on the bone feature sequence, the motion information of each bone point of the target living body is determined, wherein the motion information includes: the current position of the bone point in each bone feature vector, and the confidence of the current position, which is calculated by an OKS algorithm and used to reflect the similarity between the current position of the bone point and a preset position.

[0015] Since the user to which the target face information belongs is determined in the foregoing step, the preset position of the bone point corresponding to the user can be obtained at this time to calculate the confidence of the current position of the bone point. Thus, when calculating the activity amount of the user, the individual differences of the user are taken into account, which can effectively improve the accuracy of activity amount calculation.

[0016] Finally, based on the motion information of each skeleton point of the target living body, an activity amount of the target living body is determined, and the activity amount is used to calculate motion feature information of the target living body. The motion information of the skeleton point reflects a position change of the skeleton point in each second image in the skeleton feature sequence with a change in a posture of the target living body, and the position change of each skeleton point reflects an activity amount of the target living body to a certain extent. Therefore, the skeleton point can be taken as an activity amount calculation unit, and based on the motion information of each skeleton point of the target living body, a total amount of the change in the posture of the target living body, that is, the activity amount, is calculated. In addition, the motion feature information includes, but is not limited to, calorie, motion intensity, and heart-lung load intensity.

[0017] The above technical solution can flexibly calculate the activity amount of the user in combination with the position of the skeleton point of the user and the motion of the skeleton point in the change in the posture of the user, effectively considers individual differences of the user, and improves the accuracy and reliability of the activity amount calculation. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 A flowchart of an information processing method according to one embodiment of the present application is shown;

[0020] Figure 2 A flowchart of an information processing method according to another embodiment of the present application is shown;

[0021] Figure 3 A flowchart of an information processing method according to another embodiment of the present application is shown;

[0022] Figure 4 A flowchart of an information processing method according to another embodiment of the present application is shown;

[0023] Figure 5 A schematic diagram of an information processing process according to one embodiment of the present application is shown;

[0024] Figure 6 A block diagram of an information processing device according to one embodiment of the present application is shown;

[0025] Figure 7 A block diagram of a computer device according to one embodiment of the present application is shown;

[0026] Figure 8A block diagram of a computer device according to another embodiment of the present application is shown. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of the present application.

[0028] Embodiment One

[0029] Figure 1 A flowchart of an information processing method according to one embodiment of the present application is shown.

[0030] As Figure 1 shown, the information processing method according to one embodiment of the present application includes:

[0031] Step 102, acquiring a first image sequence, the first image sequence including consecutive frame images in a target video.

[0032] The first image sequence is a video frame of a video. Specifically, a video can be collected by any computer device capable of video collection such as a mobile phone, a wearable device, a monitoring device, etc., and a first image sequence composed of video frames in the video is taken as the basis for activity quantity calculation.

[0033] Step 104, if the first image sequence has a target living body, extracting a second image sequence in the first image sequence based on the similarity of each group of adjacent first images in the first image sequence.

[0034] The target living body refers to a predetermined object that needs to be calculated for activity quantity, such as a user of a wearable device.

[0035] In one possible design, whether the first image sequence has a living body can be detected based on a predetermined living body detection method first. The predetermined living body detection method includes but is not limited to infrared image living body detection, infrared image living body detection, RGB image living body detection, and video living body detection under the face and human category provided by Ali Cloud, etc.

[0036] Then, if the first image sequence has a living body, whether each first image of the first image sequence has a face is identified by an MTCNN algorithm. The MTCNN algorithm is a face detection and face alignment method based on deep learning, which can complete the tasks of face detection and face alignment at the same time, and has higher detection efficiency and detection accuracy.

[0037] If there is a face in each first image of the first image sequence, target face information matching the face is selected from a plurality of preset face information, and the target living body is determined as a user to which the target face information belongs. For a computer device such as a mobile phone or a wearable device, a single device can have multiple users. Therefore, it can be determined through face comparison which user of the device the target living body belongs to, so as to effectively distinguish and count information of different users.

[0038] The first image sequence includes a plurality of first images. If the similarity of a group of adjacent first images is high, it indicates that the body state of the target living body in the two first images of the group is similar, and the target living body does not perform a large motion in the time interval corresponding to the two first images of the group. Conversely, if the similarity of a group of adjacent first images is low, it indicates that the body state of the target living body in the time interval corresponding to the two first images of the group is different, and the target living body performs a large motion in the time interval corresponding to the two first images of the group, and the body state changes greatly. Therefore, the similarity of adjacent first images reflects the activity of the target living body, so the second image in which the body state of the target living body changes can be extracted from the first image sequence based on the similarity of each group of adjacent first images in the first image sequence, to form a second image sequence. That is, the second image sequence is an image in which the body state of the target living body changes, reflecting the activity of the target living body.

[0039] In a possible design, the similarity is calculated in the following manner: the foreground target of each group of adjacent first images in the first image sequence is identified through a ViBe algorithm; common feature points of the foreground target of each group of adjacent first images are determined based on scale-invariant feature transform features of the foreground target; and the similarity of each group of adjacent first images is determined based on position change information of the common feature points in the each group of adjacent first images.

[0040] The ViBe algorithm is a background difference method of background modeling first and foreground detection later, and has the advantages of small memory occupation and fast calculation speed. The scale-invariant feature transform (SIFT) feature can effectively describe the local details of the foreground target. By using the scale-invariant feature transform feature, the feature points in the foreground targets of two adjacent first images can be effectively distinguished, and the feature points of the two adjacent first images can be matched to obtain common feature points of the two adjacent first images. The position change information of the common feature points, that is, the difference between the position of the common feature points in one first image and the position of the common feature points in another first image, can represent the body state change of the target living body, and thus reflects the similarity of the two adjacent first images.

[0041] In a possible design, for two adjacent first images, a ratio of a position offset distance of each common feature point of the two to a predetermined offset distance is determined, and a mean of the ratio of all common feature points of the two is determined as a similarity of the two.

[0042] In step 106, a skeletal feature sequence is obtained by acquiring, based on a predetermined skeletal recognition algorithm, a skeletal feature vector of each second image in the second image sequence.

[0043] The change of the skeletal feature vector of each second image in the skeletal feature sequence reflects the posture change of the target living body reaching the activity amount statistical standard. Specifically, for each second image, the predetermined skeletal recognition algorithm includes: determining a plurality of skeletal points of the target living body in the second image based on a predetermined posture estimation algorithm; and generating a skeletal feature vector of the target living body based on the plurality of skeletal points, wherein the skeletal feature vectors of all the second images in the second image sequence constitute the skeletal feature sequence.

[0044] The predetermined posture estimation algorithm can be 2D posture estimation or 3D posture estimation, including but not limited to any algorithm capable of posture estimation such as top-down, bottom-up, CPM, Hourglass, etc.

[0045] In step 108, motion information of each skeletal point of the target living body is determined based on the skeletal feature sequence, wherein the motion information includes: a current position of the skeletal point in each skeletal feature vector, and a confidence of the current position, which is calculated by an OKS algorithm and used to reflect the similarity of the current position of the skeletal point and a preset position.

[0046] The OKS algorithm is an algorithm for calculating a similarity index between points, and in the present technical solution, the similarity index refers to the confidence of the current position. Specifically, a similarity matrix can be constructed by the skeletal feature vector of each second image in the second image sequence, wherein an element in the i-th row and the j-th column of the similarity matrix is the similarity of the current position of a skeletal point x in the i-th second image and a preset position of the skeletal point x in the j-th second image. The maximum similarity in the i-th row is the confidence of the skeletal point x in the i-th second image.

[0047] Since the user to which the target face information belongs is determined in the foregoing steps, the preset positions of the skeletal points corresponding to the user can be acquired to calculate the confidence of the current position of the skeletal points. Thus, when calculating the activity amount of the user, the individual difference of the user is considered, and the accuracy of the activity amount calculation can be effectively improved.

[0048] In step 110, based on the motion information of each skeleton point of the target living body, an activity amount of the target living body is determined, and the activity amount is used to calculate the motion feature information of the target living body.

[0049] The motion information of the skeleton point reflects the position change of the skeleton point in each second image in the skeleton feature sequence with the posture change of the target living body, and the position change of each skeleton point reflects the activity amount of the target living body to a certain extent. Therefore, the skeleton point can be taken as an activity amount calculation unit, and based on the motion information of each skeleton point of the target living body, the total amount of posture change, i.e., the activity amount, of the target living body is calculated.

[0050] In addition, the motion feature information includes, but is not limited to, calorie, motion intensity, and cardiopulmonary load intensity.

[0051] In a possible design, based on the motion information of each skeleton point, a motion trajectory of each skeleton point of the target living body is determined, and the length of the motion trajectory of each skeleton point is taken as the moving distance of each skeleton point; and based on the moving distance of each skeleton point, the activity amount of the target living body is determined.

[0052] Optionally, the sum of the moving distances of the skeleton points of the target living body is taken as the activity amount of the target living body.

[0053] In addition, in a possible design, after the video is collected in real time, the activity amount is calculated directly based on the image sequence of the video, and dynamic calculation of the activity amount is realized. In this calculation manner, the corresponding video can be deleted after each calculation, and the storage resource is released.

[0054] The above technical solution can flexibly calculate the activity amount of the user in combination with the position of the skeleton point of the user and the motion of the skeleton point in the posture change of the user, effectively considers the individual difference of the user, and improves the accuracy and reliability of the activity amount calculation.

[0055] Taking the intelligent family doctor service as an example, the activity amount of the user is calculated by using the above technical solution, and the motion intensity and the cardiopulmonary load intensity of the user can be effectively calculated, so that warning information can be sent in time when the motion intensity and the cardiopulmonary load intensity of the user are insufficient or exceed the standard.

[0056] Taking the APP and the small program for fitness service as an example, the activity amount of the user is calculated by using the above technical solution, so that the calorie consumption of the user can be effectively calculated and displayed in the APP or the small program, and the user can timely know the fitness condition of the user, and the user experience is improved.

[0057] Taking an insurance business as an example, the activity amount of a user is calculated by using the technical solution, so that the motion intensity and the cardiopulmonary load intensity of the user can be effectively calculated, the health management level of the user is evaluated based on the motion intensity and the cardiopulmonary load intensity of the user, and appropriate insurance services are recommended for the user according to the health management level of the user, or the premium of the user is adjusted according to the health management level of the user.

[0058] Embodiment Two

[0059] Based on the embodiment one, Figure 2 A flowchart of an information processing method according to another embodiment of the present application is shown.

[0060] As Figure 2 shown, the information processing method according to another embodiment of the present application comprises:

[0061] Step 202, a first image sequence is acquired, and the first image sequence comprises consecutive frame images in a target video.

[0062] Step 204, if there is a target living body in the first image sequence, a similarity distribution curve is generated based on the similarity of each group of adjacent first images in the first image sequence.

[0063] Step 206, adjacent first images with the similarity located at a trough of the similarity distribution curve are extracted to the second image sequence.

[0064] The horizontal axis of the coordinate system in which the similarity distribution curve is located represents time, and the vertical axis represents the similarity of adjacent first images. The smaller the posture change of the target living body in the time interval corresponding to adjacent first images, the higher the similarity of adjacent first images, and correspondingly, the greater the similarity of adjacent first images in the similarity distribution curve. In the similarity distribution curve, the greater the similarity is often distributed at the peak, and correspondingly, the smaller the similarity is often distributed at the peak.

[0065] Therefore, the adjacent first images with the similarity located at the trough of the similarity distribution curve can be directly determined as images when the body posture of the target living body changes, and added to the second image sequence.

[0066] Step 208, based on a predetermined bone recognition algorithm, a bone feature vector of each second image in the second image sequence is acquired to obtain a bone feature sequence.

[0067] Step 210, based on the bone feature sequence, motion information of each bone point of the target living body is determined, wherein the motion information comprises a current position of the bone point in each bone feature vector and a confidence degree of the current position, and the confidence degree is calculated by an OKS algorithm and is used to reflect the similarity between the current position of the bone point and a preset position.

[0068] Step 212, based on the motion information of each bone point of the target living body, determining the activity amount of the target living body, the activity amount being used to calculate the motion feature information of the target living body.

[0069] Embodiment Three

[0070] On the basis of Embodiment One, Figure 3 A flow chart of an information processing method according to still another embodiment of the application is shown.

[0071] As Figure 3 shown, the information processing method according to still another embodiment of the application comprises:

[0072] Step 302, acquiring a first image sequence, the first image sequence comprising consecutive frame images within a target video.

[0073] Step 304, if the first image sequence has a target living body, generating a similarity distribution curve based on the similarity of each group of adjacent first images in the first image sequence.

[0074] Step 306, extracting, to the second image sequence, adjacent first images whose similarity on the similarity distribution curve has a slope absolute value greater than or equal to a predetermined slope threshold value.

[0075] The horizontal axis of the coordinate system in which the similarity distribution curve lies represents time, and the vertical axis represents the similarity size of adjacent first images. For the two groups of adjacent first images, if the target living body has a posture change within the time interval between the two groups of adjacent first images, the similarity of the former group of adjacent first images is very high, and the similarity of the latter group of adjacent first images is very low. Correspondingly, the two groups of adjacent first images have a sharp drop on the similarity distribution curve. In the case that the time length between each first image is fixed, the steeper the similarity distribution curve is, the higher the slope absolute value is.

[0076] Therefore, the predetermined slope threshold value can be set, which is the lowest slope absolute value corresponding to the similarity change when the posture change of the target living body reaches the activity amount statistical standard. In this way, if the similarity of adjacent first images has a slope absolute value on the similarity distribution curve greater than or equal to the predetermined slope threshold value, it indicates that the target living body has a posture change reaching the activity amount statistical standard within the time interval before or after the adjacent first image.

[0077] At this time, the adjacent first image can be determined as the image when the body posture of the target living body changes, and added to the second image sequence.

[0078] Step 308, based on a predetermined bone recognition algorithm, acquiring the bone feature vector of each second image in the second image sequence, to obtain a bone feature sequence.

[0079] In step 310, motion information of each skeletal point of the target living body is determined based on the skeletal feature sequence, wherein the motion information comprises a current position of the skeletal point in each skeletal feature vector and a confidence of the current position, and the confidence is calculated by the OKS algorithm and is used to reflect a similarity between the current position and a preset position of the skeletal point.

[0080] In step 312, an activity amount of the target living body is determined based on the motion information of each skeletal point of the target living body, and the activity amount is used to calculate motion feature information of the target living body.

[0081] In addition, in a possible design, the adjacent first image with the similarity less than the predetermined similarity threshold value can be directly extracted to the second image sequence, so as to achieve the purpose of quickly extracting the second image sequence.

[0082] Embodiment Four

[0083] Figure 4 A flowchart of an information processing method according to yet another embodiment of the application is shown.

[0084] As shown in Figure 4 An information processing method according to yet another embodiment of the application comprises:

[0085] In step 402, a first image sequence is acquired, and the first image sequence comprises continuous frame images in a target video.

[0086] In step 404, if the target living body exists in the first image sequence, a second image sequence is extracted from the first image sequence based on a similarity of each group of adjacent first images in the first image sequence.

[0087] In step 406, a skeletal feature vector of each second image in the second image sequence is acquired based on a predetermined skeletal recognition algorithm, and a skeletal feature sequence is obtained.

[0088] In step 408, motion information of each skeletal point of the target living body is determined based on the skeletal feature sequence, wherein the motion information comprises a current position of the skeletal point in each skeletal feature vector and a confidence of the current position, and the confidence is calculated by the OKS algorithm and is used to reflect a similarity between the current position and a preset position of the skeletal point.

[0089] In step 410, motion information of each limb part of the target living body is determined based on the motion information of each skeletal point of the target living body.

[0090] Each of the body parts of the target living body is composed of a plurality of skeletal points, and the motion information of each of the body parts includes motion information of the plurality of skeletal points that compose the body part.

[0091] Further, the body parts include limbs and a trunk. The motion of the target living body is the posture change of the limbs and the trunk, and therefore the skeletal points can be combined to obtain the skeletal point combination of the limbs and the trunk respectively. Correspondingly, based on the motion information of each of the skeletal points, the motion information of each of the body parts of the target living body, such as the limbs and the trunk, is determined.

[0092] In step 412, the motion trajectory of each of the body parts is determined based on the motion information of each of the body parts, and the length of the motion trajectory of each of the body parts is taken as the moving distance of each of the body parts.

[0093] In one possible design, for each of the body parts of the target living body, the average of the current position coordinates of all the skeletal points of the body part in each of the second images is determined as the coordinates of the virtual position of the body part in each of the second images, and the line connecting the virtual positions of the body part in all the second images is determined as the motion trajectory of the body part.

[0094] In other words, in a single second image, the virtual position corresponding to a single body part is determined based on the coordinates of the plurality of skeletal points of the single body part. Further, the change of the virtual positions in the plurality of second images is equivalent to the motion route of the single body part.

[0095] In another possible design, for each of the body parts of the target living body, the motion trajectory of each of the skeletal points of the body part in the sequence of second images is determined, and the motion trajectory of each of the skeletal points of the body part is fitted into the motion trajectory of the body part based on a predetermined fitting manner.

[0096] In this way, the motion trajectories of the plurality of skeletal points of the body part can be directly fitted to obtain the motion trajectory of the body part. The predetermined fitting manner includes but is not limited to the least square method, the spline interpolation method, and any other curve fitting manner.

[0097] In step 414, the activity amount of the target living body is determined based on the moving distances of each of the body parts.

[0098] The moving distances of all the body parts of the target living body reflect the posture change of the target living body, i.e., the activity amount of the target living body.

[0099] In a possible design, a corrected movement distance of each limb part is determined based on the movement distance of each limb part and the activity weight of each limb part; and a sum of the corrected movement distances of each limb part of the target living body is determined as the activity amount of the target living body. In other words, the movement distances of each limb part are weighted and summed to obtain the activity amount of the target living body.

[0100] In another possible design, for each skeleton point of each limb part, a first mean value of the confidence corresponding to the skeleton point is determined; for each limb part, a second mean value is determined based on the first mean values corresponding to all skeleton points of the limb part; a corrected activity weight of the limb part is determined based on the second mean value and a mean value of the activity weights of the limb part; a corrected movement distance of each limb part is determined based on the movement distance of each limb part and the corrected activity weight of each limb part; and a sum of the corrected movement distances of each limb part of the target living body is determined as the activity amount of the target living body.

[0101] The confidence described herein, i.e., the confidence in the motion information of the skeleton point, is calculated by the OKS algorithm and is used to reflect the similarity between the current position of the skeleton point and the preset position. For a single skeleton point, the mean value of the confidence corresponding to the skeleton point in multiple second images, i.e., the first mean value, reflects the overall confidence of the skeleton point. For a single limb part, the mean value of the overall confidence of all skeleton points of the single limb part, i.e., the second mean value, reflects the overall confidence of the posture change of the single limb part.

[0102] In a possible design, the second mean value can be directly used as the activity weight of the limb part.

[0103] In another possible design, when the activity weight of the limb part is calculated for the first time, the second mean value is used as the activity weight of the limb part. When the activity weight of the limb part is adjusted subsequently, the second mean value is used as a correction parameter, and a corrected activity weight of the limb part is determined based on the second mean value and a mean value of the activity weights of the limb part. In this way, the activity weight of the limb part can be dynamically adjusted based on the real-time posture change of the target living body, thereby facilitating the improvement of the calculation accuracy of the activity amount.

[0104] Finally, the movement distances of each limb part are weighted and summed to obtain the activity amount of the target living body, and the weight of the weighting is the corrected activity weight of each limb part.

[0105] The above technical solutions introduce the limb part on the basis of the skeleton point, to more intuitively and effectively describe the posture change of the target living body, and dynamically correct the activity weight of the limb part, thereby greatly improving the accuracy of the calculation of the activity amount, providing a reliable basis for the calculation of the health parameters such as calories, and improving the user experience of the health services in digital medicine.

[0106] Figure 5 A schematic diagram of an information processing process according to an embodiment of the present application is shown.

[0107] As shown in the data acquisition phase, it includes: Figure 5

[0108] Step 502, video stream acquisition.

[0109] Video stream acquisition is a special way of data acquisition, mainly to sample, quantize, etc. Video information output by devices capable of collecting video such as cameras, video recorders, televisions, mobile phones, wearable devices, etc. into video stream picture data.

[0110] Step 504, live detection.

[0111] After obtaining the video stream picture data, live detection is needed to determine whether the biological body in the video stream picture data is a real live body or a fake attack.

[0112] Step 506, face acquisition.

[0113] After obtaining the video stream picture data, face detection is performed frame by frame through the MTCNN algorithm. It should be noted that the order of the three steps of face acquisition, live detection and skeletal feature acquisition described below is not limited.

[0114] Step 508, skeletal feature acquisition.

[0115] After obtaining the video stream picture data, skeletal feature acquisition is performed frame by frame using the OKS algorithm, and the similarity between the detected human skeletal key point position and the real annotation is scored.

[0116] Next, enter the data analysis phase.

[0117] Step 510, key frame optimization algorithm.

[0118] Specifically, the ViBe algorithm is used to detect the foreground target in the image sequence of the video stream picture data, extract the scale invariant feature transform feature of the foreground target, and match the feature points between adjacent frames. Then, the similarity of adjacent frames is calculated. Finally, according to the similarity of adjacent frames, the key frame is identified.

[0119] This technical solution can solve the problem of missing and wrong selection in key frame extraction of motion videos. Compared with the algorithm based on SIFT distribution histogram, it has higher precision and recall rate. Therefore, this technical solution has a more excellent detection effect for identifying key frames containing key actions in motion videos.

[0120] ​Step 512, face comparison algorithm.

[0121] After performing face detection on each frame of the image sequence in the video stream, the faces are compared with a preset face database to identify the user to whom the face in the video stream belongs.

[0122] Step 514, skeleton recognition algorithm.

[0123] For each frame, the human skeleton sequence is represented using 2D or 3D coordinates. Skeletal point-based motion recognition is achieved by concatenating all joint vectors in each frame into a single feature vector. Then, a spatiotemporal convolutional map is used to form a multi-layered skeleton sequence.

[0124] The spatiotemporal skeletal graph G = (V, E) is constructed, where the number of frames in the video stream image sequence is T, the number of skeletal points is N, and the node matrix set V = {vti | t = 1, ..., T, i = 1, ..., N} includes all keypoints on the skeletal sequence. When ST-GCN is used as input, the feature vector F(vti) of the t-th frame and the i-th skeletal point at the keypoint is... ti The skeletal structure consists of the coordinates and confidence level of each skeletal point. This allows for the construction of a spatiotemporal graph of the skeletal sequence, facilitating skeletal recognition.

[0125] Step 516, Spatial Action Recognition Algorithm.

[0126] To achieve motion recognition, it is not possible to do so with a single frame. Therefore, it is also necessary to collect the correlation between frames to represent the temporal relationship between corresponding skeletal points of the human body.

[0127] Within each frame, a spatial diagram is constructed according to the natural skeletal connections of the human body. This connection is established based on the natural structure, without any manual design.

[0128] Furthermore, this spatial graph supports use with varying numbers of skeleton point sets. For example, in the Kinect dataset, we obtained 17 skeleton points for 2D pose estimation using the OpenPose tool, with the intra-frame linking of skeleton points as follows: Es = {v ti v tj |(i,j)∈H}, where H represents the set of human skeleton points, and (i,j) are the coordinates of the skeleton points. The link between different frames is: EF={v ti v (t+1)i Therefore, each edge in EF represents the trajectory of a specific limb part (such as the forearm) over time, thus allowing us to determine the user's specific movement trajectory.

[0129] Finally, we move on to the data presentation stage.

[0130] Step 518, substitute the user basic information and the calorie calculation formula.

[0131] The user data is collected by a user terminal (such as a WeChat applet providing fitness services, a body measurement instrument, etc.), including height, age, weight, gender, activity amount, etc. The activity amount is calculated based on the specific motion trajectory of the user.

[0132] Further, the user data is substituted into the calorie calculation formula. Optionally, the calorie calculation formula is:

[0133] Male calorie = [66+1.38*weight(kg)+5*height(cm)-6.8*age]*activity amount;

[0134] Female calorie = [6.55+9.6*weight(kg)+1.9*height(cm)-4.7*age]*activity amount.

[0135] Step 520, output the calorie to the user terminal.

[0136] The calculation result of the calorie is displayed on the user terminal, such as a WeChat applet providing fitness services, a body measurement instrument, etc., so that the user can know the calorie consumption at the first time.

[0137] In summary, through a calorie counting method based on face recognition of users and combined with a bone action algorithm, the motion health of the fitness population can be promoted, the motion level can be effectively grasped, and the fitness experience of the user can be improved.

[0138] Figure 6 A block diagram of an information processing apparatus according to one embodiment of the present application is shown.

[0139] As shown in Figure 6 , the information processing apparatus 600 according to one embodiment of the present application includes:

[0140] A first image sequence acquisition unit 602 is configured to acquire a first image sequence, the first image sequence including consecutive frame images within a target video;

[0141] A second image sequence acquisition unit 604 is configured to, if the first image sequence has a target living body, extract a second image sequence from the first image sequence based on the similarity of each group of adjacent first images in the first image sequence;

[0142] A bone feature sequence acquisition unit 606 is configured to acquire a bone feature vector of each second image in the second image sequence based on a predetermined bone recognition algorithm, to obtain a bone feature sequence;

[0143] The skeleton point motion information determination unit 608 is configured to determine motion information of each skeleton point of the target living body based on the skeleton feature sequence, wherein the motion information comprises a current position of the skeleton point in each skeleton feature vector and a confidence of the current position, and the confidence is calculated by an OKS algorithm and is used to reflect a similarity between the current position and a preset position.

[0144] The activity amount determination unit 610 is configured to determine an activity amount of the target living body based on the motion information of each skeleton point of the target living body, and the activity amount is used to calculate motion feature information of the target living body.

[0145] The information processing apparatus 600 further comprises:

[0146] The living body detection unit is configured to detect whether there is a living body in the first image sequence based on a predetermined living body detection manner before extracting the second image sequence from the first image sequence.

[0147] The face recognition unit is configured to recognize whether there is a face in each first image of the first image sequence by using an MTCNN algorithm if there is a living body in the first image sequence.

[0148] The face matching unit is configured to select target face information matched with the face from a plurality of preset face information if there is a face in each first image of the first image sequence, and determine the target living body as a user to which the target face information belongs.

[0149] The information processing apparatus 600 further comprises:

[0150] The foreground target recognition unit is configured to recognize a foreground target of each group of adjacent first images in the first image sequence by using a ViBe algorithm before extracting the second image sequence from the first image sequence.

[0151] The common feature point extraction unit is configured to determine a common feature point of the foreground target of each group of adjacent first images based on a scale-invariant feature transform feature of the foreground target.

[0152] The similarity determination unit is configured to determine a similarity of each group of adjacent first images based on position change information of the common feature point in the each group of adjacent first images.

[0153] The second image sequence acquisition unit 604 is configured to:

[0154] generate a similarity distribution curve based on the similarity of each set of adjacent first images in the first image sequence; extract adjacent first images with the similarity located at a trough of the similarity distribution curve to the second image sequence; or extract adjacent first images with an absolute value of a slope of the similarity on the similarity distribution curve greater than or equal to a predetermined slope threshold to the second image sequence; or extract adjacent first images with the similarity less than a predetermined similarity threshold to the second image sequence.

[0155] The skeleton feature sequence acquisition unit 606 is configured to:

[0156] For each of the second images, determine a plurality of skeleton points of the target living body in the second image based on a predetermined pose estimation algorithm; and generate a skeleton feature vector of the target living body based on the plurality of skeleton points, wherein the skeleton feature vectors of all the second images in the second image sequence constitute the skeleton feature sequence.

[0157] The activity amount determination unit 610 is configured to:

[0158] Determine a motion trajectory of each skeleton point of the target living body based on the motion information of the each skeleton point, and take a length of the motion trajectory of the each skeleton point as a movement distance of the each skeleton point; and determine an activity amount of the target living body based on the movement distance of the each skeleton point.

[0159] The activity amount determination unit 610 includes:

[0160] A limb part motion information determination unit configured to determine motion information of each limb part of the target living body based on the motion information of each skeleton point of the target living body, wherein each limb part of the target living body is composed of a plurality of skeleton points, and the motion information of the each limb part includes motion information of the plurality of skeleton points composing the limb part;

[0161] A limb part motion trajectory determination unit configured to determine a motion trajectory of the each limb part based on the motion information of the each limb part, and take a length of the motion trajectory of the each limb part as a movement distance of the each limb part;

[0162] The activity amount determination unit 610 is configured to:

[0163] Determine an activity amount of the target living body based on the movement distance of the each limb part.

[0164] The limb part motion trajectory determination unit is configured to:

[0165] For each limb part of the target living body, a mean value of current position coordinates of all skeletal points of the limb part in each of the second images is determined as a coordinate of a virtual position of the limb part in each of the second images, and a line connecting the virtual positions of the limb part in all the second images is determined as a motion trajectory of the limb part.

[0166] The limb part motion trajectory determination unit is configured to:

[0167] For each limb part of the target living body, a motion trajectory of each skeletal point of the limb part in the sequence of second images is determined, and the motion trajectories of each skeletal point of the limb part are fitted into a motion trajectory of the limb part based on a predetermined fitting manner.

[0168] The activity amount determination unit 610 is configured to:

[0169] Based on the moving distance of each limb part and the activity weight of each limb part, a corrected moving distance of each limb part is determined, and a sum of the corrected moving distances of each limb part of the target living body is determined as an activity amount of the target living body.

[0170] The activity amount determination unit 610 is configured to:

[0171] For each skeletal point of each limb part, a first mean value of the confidence corresponding to the skeletal point is determined, and for each limb part, a second mean value is determined based on the first mean values corresponding to all skeletal points of the limb part. The second mean value and a mean value of the activity weights of the limb part are determined as a corrected activity weight of the limb part. Based on the moving distance of each limb part and the corrected activity weight of each limb part, a corrected moving distance of each limb part is determined, and a sum of the corrected moving distances of each limb part of the target living body is determined as an activity amount of the target living body.

[0172] The information processing apparatus 600 uses the scheme of any one of the above embodiments, and therefore has all the technical effects described above, which will not be repeated here.

[0173] In addition, in one embodiment, the present application provides a computer device, which can be a server, and an internal structure diagram of the computer device can be as shown in Figure 7As shown in the figure. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes non-volatile and / or volatile storage media, internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with the external client through the network connection. The computer program is executed by the processor to implement the information processing method described in any of the above embodiments.

[0174] In one embodiment, the present application also provides a computer device, which can be a client, and its internal structure diagram can be as shown in the figure. Figure 8 As shown in the figure. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes non-volatile storage media, internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with the external server through the network connection. The computer program is executed by the processor to implement the information processing method described in any of the above embodiments.

[0175] Any computer device described in the above embodiments of the present application exists in various forms, including but not limited to:

[0176] (1) Mobile communication device: This kind of device is characterized by mobile communication function, and the main target is to provide voice and data communication. This kind of terminal includes: smart phone (such as iPhone), multimedia phone, functional phone, and low-end phone, etc.

[0177] (2) Ultra-mobile personal computer device: This kind of device belongs to the category of personal computer, has computing and processing function, and generally also has mobile Internet characteristics. This kind of terminal includes: PDA, MID and UMPC device, etc., such as iPad.

[0178] (3) Portable entertainment device: This kind of device can display and play multimedia content. This kind of device includes: audio and video player (such as iPod), handheld game console, electronic book, and smart toy, wearable device and portable car navigation device.

[0179] (4) Server: a device providing computing services, the configuration of the server includes a processor, a hard disk, a memory, a system bus, etc., the server is similar to a general computer architecture, but since it needs to provide high-reliable services, it has higher requirements in processing capability, stability, reliability, security, scalability, manageability, etc.

[0180] (5) Other electronic devices with data interaction function.

[0181] In addition, the embodiment of the present application provides a computer readable storage medium storing computer executable instructions, the computer executable instructions are used for executing the following steps:

[0182] Obtaining a first image sequence, the first image sequence includes consecutive frame images in a target video;

[0183] If the first image sequence has a target living body, based on the similarity of each group of adjacent first images in the first image sequence, a second image sequence is extracted in the first image sequence;

[0184] Based on a predetermined bone recognition algorithm, a bone feature vector of each second image in the second image sequence is obtained, and a bone feature sequence is obtained;

[0185] Based on the bone feature sequence, motion information of each bone point of the target living body is determined, wherein the motion information includes a current position of the bone point in each bone feature vector and a confidence of the current position, the confidence is calculated by an OKS algorithm and is used to reflect the similarity between the current position of the bone point and a preset position;

[0186] Based on the motion information of each bone point of the target living body, an activity amount of the target living body is determined, and the activity amount is used to calculate motion feature information of the target living body.

[0187] It should be noted that the functions or steps that the computer readable storage medium or the computer device can achieve described above can be referred to the related description in the foregoing method embodiment, and here will not be described one by one to avoid repetition.

[0188] The technical scheme of the present application is described in detail above in combination with the drawings, through the technical scheme of the present application, the activity amount of the user can be flexibly calculated by combining the bone point position of the user itself with the bone point motion in the user posture change, the individual difference of the user is effectively considered, and the accuracy and reliability of the activity amount calculation are improved.

[0189] It should be understood that, although the terms first, second, etc. can be used herein to describe various images, these images should not be limited to these terms. These terms are only used to distinguish one image from another. For example, a first image could be termed a second image without departing from the scope of embodiments of the application, and, similarly, a second image could be termed a first image.

[0190] The word "if" can be interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if (a stated condition or event) is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)," depending on the context.

[0191] The terminology used in the embodiments of the application, and the appended claims, is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the description of the embodiments of the application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0192] In several embodiments provided in the application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are merely illustrative, and the division of the units is only a logical function division. In actual implementation, another division manner can be adopted, for example, a plurality of units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the units or components shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0193] In addition, each functional unit in the various embodiments of the application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or hardware plus software function unit.

[0194] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0195] The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. An information processing method, characterized in that, include: Obtain a first image sequence, the first image sequence comprising consecutive frame images within the target video; If the first image sequence contains a target living body, a second image sequence is extracted from the first image sequence based on the similarity of each group of adjacent first images in the first image sequence; Based on a predetermined skeleton recognition algorithm, the skeleton feature vector of each second image in the second image sequence is obtained to obtain a skeleton feature sequence; Based on the skeletal feature sequence, motion information of each skeletal point of the target living body is determined, wherein the motion information includes: the current position of the skeletal point in each skeletal feature vector, and the confidence level of the current position, which is calculated by the OKS algorithm and is used to reflect the similarity between the current position of the skeletal point and a preset position; Based on the motion information of each skeletal point of the target living body, the activity level of the target living body is determined, and the activity level is used to calculate the motion characteristic information of the target living body.

2. The information processing method according to claim 1, characterized in that, Before extracting the second image sequence from the first image sequence, the process further includes: Based on a predetermined liveness detection method, detect whether there is a live body in the first image sequence; If there is a live subject in the first image sequence, the MTCNN algorithm is used to identify whether there is a human face in each first image of the first image sequence. If each first image in the first image sequence contains a human face, select the target human face information that matches the human face from multiple preset human face information, and determine the target liveness as the user to which the target human face information belongs.

3. The information processing method according to claim 1, characterized in that, Before extracting the second image sequence from the first image sequence, the process further includes: The ViBe algorithm is used to identify the foreground target in each group of adjacent first images in the first image sequence. Based on the scale-invariant feature transformation features of the foreground target, the common feature points of the foreground target in each group of adjacent first images are determined; The similarity of each group of adjacent first images is determined based on the position change information of the common feature points in each group of adjacent first images.

4. The information processing method according to claim 1 or 2, characterized in that, The step of extracting a second image sequence from the first image sequence based on the similarity of each group of adjacent first images in the first image sequence includes: A similarity distribution curve is generated based on the similarity of each group of adjacent first images in the first image sequence; Extract the adjacent first images whose similarity falls at the trough of the similarity distribution curve into the second image sequence; or The adjacent first images whose slope absolute value on the similarity distribution curve is greater than or equal to a predetermined slope threshold are extracted into the second image sequence; or The adjacent first images with a similarity less than a predetermined similarity threshold are extracted into the second image sequence.

5. The information processing method according to claim 1, characterized in that, The predetermined skeleton recognition algorithm obtains the skeleton feature vector of each second image in the second image sequence to obtain a skeleton feature sequence, including: For each of the second images, multiple skeletal points of the target living organism in the second image are determined based on a predetermined pose estimation algorithm; Based on the multiple skeletal points, a skeletal feature vector of the target living organism is generated. The skeletal feature sequence is composed of the skeletal feature vectors of all the second images in the second image sequence.

6. The information processing method according to claim 5, characterized in that, Determining the activity level of the target living organism based on the motion information of each skeletal point includes: Based on the motion information of each bone point, the motion trajectory of each bone point of the target living body is determined, and the length of the motion trajectory of each bone point is taken as the moving distance of each bone point. The amount of activity of the target living organism is determined based on the distance moved by each skeletal point.

7. The information processing method according to claim 5, characterized in that, Determining the activity level of the target living organism based on the motion information of each skeletal point includes: Based on the motion information of each skeletal point of the target living body, the motion information of each limb of the target living body is determined, wherein each limb of the target living body is composed of multiple skeletal points, and the motion information of each limb includes the motion information of the multiple skeletal points that make up the limb. Based on the motion information of each limb, the motion trajectory of each limb is determined, and the length of the motion trajectory of each limb is taken as the movement distance of each limb. The amount of activity of the target living body is determined based on the distance of movement of each limb.

8. The information processing method according to claim 7, characterized in that, Determining the movement trajectory of each limb based on the movement information of each limb includes: For each limb of the target living organism, the average current position coordinates of all skeletal points of the limb in each second image are determined, and used as the coordinates of the virtual position of the limb in each second image; The line connecting the virtual positions of the limb in all the second images is determined as the motion trajectory of the limb.

9. The information processing method according to claim 7, characterized in that, Determining the movement trajectory of each limb based on the movement information of each limb includes: For each limb of the target living organism, determine the motion trajectory of each skeletal point of the limb within the second image sequence; Based on a predetermined fitting method, the motion trajectory of each skeletal point of the limb is fitted to the motion trajectory of the limb.

10. The information processing method according to claim 7, characterized in that, Determining the activity level of the target living body based on the movement distance of each limb includes: Based on the movement distance of each limb and the activity weight of each limb, the corrected movement distance of each limb is determined; The sum of the corrected movement distances of each limb of the target living organism is determined as the activity level of the target living organism.

11. The information processing method according to claim 7, characterized in that, Determining the activity level of the target living body based on the movement distance of each limb includes: For each skeletal point of each limb, determine the first mean of the confidence level corresponding to the skeletal point; For each limb segment, a second mean is determined based on a first mean corresponding to all skeletal points of the limb segment; The second mean and the mean of the activity weights of the limbs are determined as the corrected activity weights of the limbs; The corrected movement distance of each limb is determined based on the movement distance of each limb and the corrected activity weight of each limb; The sum of the corrected movement distances of each limb of the target living organism is determined as the activity level of the target living organism.

12. An information processing device, characterized in that, include: The first image sequence acquisition unit is used to acquire a first image sequence, which includes consecutive frame images within the target video. The second image sequence acquisition unit is used to extract a second image sequence from the first image sequence if there is a target living body in the first image sequence, based on the similarity of each group of adjacent first images in the first image sequence. The skeletal feature sequence acquisition unit is used to acquire the skeletal feature vector of each second image in the second image sequence based on a predetermined skeletal recognition algorithm, thereby obtaining a skeletal feature sequence. A skeletal point motion information determination unit is used to determine the motion information of each skeletal point of the target living body based on the skeletal feature sequence. The motion information includes: the current position of the skeletal point in each skeletal feature vector, and the confidence level of the current position. The confidence level is calculated by the OKS algorithm and is used to reflect the similarity between the current position of the skeletal point and a preset position. The activity level determination unit is used to determine the activity level of the target living body based on the motion information of each skeletal point of the target living body. The activity level is used to calculate the motion characteristic information of the target living body.

13. A computer device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, the instructions being configured to perform the method according to any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that, The device stores computer-executable instructions for performing the method as described in any one of claims 1 to 11.

Images