A target movement triggering method and system based on human body image comparison

Abstract

The present invention discloses a method and system for triggering target movement based on human body image comparison, the method includes the following steps: obtaining the position feature point P1 of the current frame human body image; the position feature point P1 is obtained based on the intersection of the scanning line L and the boundary contour line of the target human body in the current frame human body image; obtaining the position feature point P2 of the comparison frame human body image; based on the position feature point P1 and the position feature point P2, calculating the movement amplitude S of the current frame human body image; judging whether the movement amplitude S of the current frame human body image meets the preset conditions, if so, triggering the prop movement instruction; after the prop movement instruction is triggered, changing the visual relative position of the target prop and / or target human body in the target picture based on the same rule. The present invention can accurately and real-time identify the action signal that the user wants to express in a real-person game, and then drive the target prop and / or target human body to move according to the preset rules.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a target movement triggering method and system based on human body image comparison. Background Art

[0002] Somatosensory games have become quite popular. Existing technology can only obtain human motion parameters by photographing the actual human body, and adjust the motion and actions of the virtual characters displayed in the game screen accordingly. If the real human body is directly displayed in the game screen to replace the virtual character, the user can see the good integration of his or her real image and real action with the game plot, which will greatly enhance the fun and sense of accomplishment of the game.

[0003] In traditional games, the range and scale of activities of virtual characters are all controlled by the game software to meet the plot requirements of the game screen. However, after the real human body image is placed in the game screen, the activities and movements of the real human body in reality are not controlled by the game software, and there are differences between the diverse shooting environments on the user side and the shooting and production environments of the game screen, such as camera device parameters, the distance between the person being photographed and the camera device, etc. This will make it difficult for users to complete the movements that meet the needs of the game plot or cause uncomfortable visual effects.

[0004] For example, in the game plot, the character should only walk on the road in the middle of the screen and should not walk into the river on the left side of the screen. When the user moves to the left relative to the camera in reality, an uncomfortable visual effect may occur in which the user's body image appears in the river in the game screen. Therefore, it is necessary to introduce target props and trigger the user's body image and target props to perform the overall movement preset by the game software through specific signals to obtain good picture effects and user experience.

[0005] However, when playing live games, it is impossible for users to hold mobile phones or keyboards to send trigger signals for the movement of props. Therefore, it is necessary to trigger screen events based on the user's specific actions. In view of the complexity and randomness of human body movements, such specific action signals must be easy to identify and cannot be misidentified. However, due to the complexity of human activities, there is currently a lack of mature technology for obtaining such action signals. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a target movement triggering method and system based on human body image comparison in view of the deficiencies in the above-mentioned prior art, which can accurately and in real time identify the action signals that the user wants to express in a real-life game in the complex and random action sequences of the human body, and then drive the target props and / or the target human body to move as a whole according to a preset trajectory. At the same time, it avoids the interference of other irrelevant actions of the human body, realizes the user's body feeling of stable action expectation and real-time action response, and reduces the user's sense of restraint when playing games. In addition, because the recognition is based on the comparison of human body planar images, there is no need for a high-cost depth information camera device. The present invention can also be used in the fields of fun selfies, fitness, etc.

[0007] In order to solve the above technical problems, the first aspect of the present invention discloses a target movement triggering method based on human body image comparison, comprising the following steps:

[0008] Acquire a position feature point P1 of the human body image of the current frame; the position feature point P1 is obtained based on the intersection of the scan line L and the boundary contour line of the target human body in the human body image of the current frame;

[0009] Acquire a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scan line L and the boundary contour line of the target human body in the comparison frame human body image;

[0010] The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image;

[0011] Based on the position feature point P1 and the position feature point P2, the movement amplitude S of the human body image in the current frame is calculated;

[0012] Determine whether the movement amplitude S of the human body image in the current frame meets the preset conditions. If so, trigger the prop movement instruction;

[0013] After the prop movement instruction is triggered, the visual relative position of the target prop and / or target human body in the target picture is changed based on the same rule.

[0014] Furthermore, the comparison frame human body image is a frame human body image among K frames human body images before the current frame human body image, or the comparison frame human body image is a frame human body image among several frames human body images within a time length K before the current frame human body image.

[0015] Furthermore, when calculating the current movement amplitude S, the position feature point P1 and the position feature point P2 are projected onto the preset trajectory line M, and the movement amplitude S is generated based on the segment of the preset trajectory line M between the projection point P1' of the position feature point P1 and the projection point P2' of the position feature point P2.

[0016] Furthermore, the scanning line L and the preset trajectory line M are parallel.

[0017] Furthermore, the movement amplitude S includes a movement direction T and a movement distance D.

[0018] Furthermore, the moving distance D is generated based on the ratio of the relative position parameter between the position feature point P1 and the position feature point P2 to the size of the specific part of the human body.

[0019] Furthermore, the size of the specific human body part is a relative position parameter between the position feature point P1 and the human body support point U1 on the current frame human body image, and the human body support point U1 is a recognizable feature point corresponding to the human body support part on the target human body.

[0020] Furthermore, the moving distance D is generated based on the angle C between the line O1 and the line O2, the line O1 refers to the line connecting the position feature point P1 and the human body support point U1 on the current frame human body image, and the line O2 refers to the line connecting the position feature point P2 and the human body support point U2 on the comparison frame human body image, and the human body support point U1 and the human body support point U2 are both identifiable feature points corresponding to the human body support parts on the target human body.

[0021] Furthermore, the method further includes executing the following steps before triggering the prop movement instruction:

[0022] Step 1, mark the current frame human body image as the comparison frame human body image;

[0023] Step 2, read the next frame of human body image as the new current frame of human body image;

[0024] Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image;

[0025] Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame;

[0026] If the moving distance D>preset threshold a, and the moving direction T is positive, then mark the current frame of human body image as the same direction frame;

[0027] If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame;

[0028] Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

[0029] The second aspect of the present invention also discloses a target movement triggering system based on human body image comparison, comprising a first acquisition module, a second acquisition module, a calculation module, a judgment module and an execution module;

[0030] The first acquisition module is used to acquire a position feature point P1 of a human body image of a current frame; the position feature point P1 is obtained based on an intersection of a scanning line L and a boundary contour line of a target human body in the human body image of the current frame;

[0031] The second acquisition module is used to acquire a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scanning line L and the boundary contour line of the target human body in the comparison frame human body image;

[0032] The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image;

[0033] The calculation module is used to calculate the movement amplitude S of the human body image of the current frame based on the position feature point P1 and the position feature point P2;

[0034] The judging module is used to judge whether the movement amplitude S of the human body image in the current frame meets the preset conditions, and if so, trigger the prop movement instruction;

[0035] The execution module is used to change the visual relative position of the target prop and / or the target human body in the target picture based on the same rule after triggering the prop movement instruction.

[0036] Furthermore, the comparison frame human body image is a frame human body image among K frames human body images before the current frame human body image, or the comparison frame human body image is a frame human body image among several frames human body images within a time length K before the current frame human body image.

[0037] Further, the movement amplitude S includes a movement direction T and a movement distance D;

[0038] The judgment module performs the following steps before triggering the prop movement instruction:

[0039] Step 1, mark the current frame human body image as the comparison frame human body image;

[0040] Step 2, read the next frame of human body image as the new current frame of human body image;

[0041] Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image;

[0042] Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame;

[0043] If the moving distance D>preset threshold a, and the moving direction T is positive, then mark the current frame of human body image as the same direction frame;

[0044] If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame;

[0045] Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

[0046] Compared with the prior art, the present invention has the following advantages:

[0047] 1. The present invention can drive the target props and the target human body to move as a whole through actions that conform to the daily perception of the human body, thereby achieving a good real-life game plot effect.

[0048] 2. Based on the intersection of the scanning line L and the boundary contour line of the target human body in the current frame human body image, the position feature point P1 is obtained to accurately identify the movement amplitude of the target part with a small depth activity range on the human body plane image, thereby triggering the prop movement instruction at a low cost.

[0049] 3. By identifying the movement amplitude and speed of the target part (which can be calculated based on the time difference between the comparison frame human body image and the current frame human body image), the interference of other irrelevant human body movements is avoided and the user's sense of restraint is reduced, so as to achieve a user's sense of stable movement expectation and real-time movement response.

[0050] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The present invention is a flow chart of the method.

[0052] Figure 2 Schematic diagram of driving the target props and target human body to move.

[0053] Figure 3 Schematic diagram of the angle between line O1 and line O2.

[0054] Figure 4 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0055] Example 1

[0056] like Figure 1 As shown, a target movement triggering method based on human body image comparison includes the following steps:

[0057] Step 1, obtaining a position feature point P1 of a human body image of a current frame; the position feature point P1 is obtained based on the intersection of a scan line L and a boundary contour line of a target human body in the human body image of the current frame;

[0058] Step 2, obtaining a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scan line L and the boundary contour line of the target human body in the comparison frame human body image;

[0059] The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image;

[0060] It should be noted that the current frame human body image can be obtained by shooting, for example, in actual operation, the user shoots himself with a camera device above the TV, and observes the fusion of his own human body image and the target image on the TV screen and performs corresponding actions to generate a human body image; the target part is a human body part such as a joint part corresponding to the hip / knee / shoulder, and the position feature point P1 / P2 is the intersection of the scanning line L and the human body boundary contour line corresponding to the human body part on the target human body, such as the intersection of the scanning line L passing through the left hip joint part and the outer boundary contour line of the left hip joint part, or the intersection of the scanning line L passing through the right hip joint part and the outer boundary contour line of the right hip joint part; it should be noted that the order of step 1 and step 2 can be interchanged;

[0061] Preferably, the position feature point P1 is obtained by taking the intersection of a scan line L passing through the target part of the target human body in the current frame human body image and the boundary contour line of the target human body in the current frame human body image, and the position feature point P2 is obtained by taking the intersection of a scan line L passing through the target part of the target human body in the comparison frame human body image and the boundary contour line of the target human body in the comparison frame human body image; that is, the scan line L corresponding to the comparison frame human body image and the scan line L corresponding to the current frame human body image are not the same scan line.

[0062] The comparison frame human body image is one of several frames of historical human body images before the current frame human body image;

[0063] It should be noted that in actual applications, the number of comparison frame human body images or current frame human body images can be multiple. For example, multiple comparison frame human body images are compared with the current frame human body image to calculate the following movement amplitude S, and then the average of the multiple movement amplitudes S is taken to reduce possible information collection errors of the camera device.

[0064] It should be noted that the position feature points P1 and P2 can also be obtained based on the intersection of the same scanning line L passing through the target part of the target human body in the current frame or the comparison frame human body image and the boundary contour line of the target human body in the current frame and the comparison frame, but the accuracy will be poor. In addition, the comparison frame human body image and its position feature point P2 can also be obtained by other means, such as pre-marking a preset human body image, to meet the needs of different game plots.

[0065] Step 3, based on the position feature point P1 and the position feature point P2, calculate the movement amplitude S of the human body image of the current frame;

[0066] Step 4: determine whether the movement amplitude S of the human body image in the current frame meets the preset conditions. If so, trigger the prop movement instruction;

[0067] Step 5: After the prop movement instruction is triggered, the visual relative position of the target prop and / or target human body in the target picture is changed based on the same rule.

[0068] By adopting the above technical solution, the following technical effects are achieved:

[0069] In order to solve the problem that in real-life games, users drive the target props and their own human images to move in the target screen according to the same rules through their own actions, it is necessary to introduce reasonable human action parts such as hips / knees / shoulders, etc., which can not only conform to the daily perception and common sense of the human body when driving the props to move to obtain a good body feeling (for example, the target prop scooter is controlled to move together with the target human body standing on the scooter through the movement and shaking of the target human body's knees, for example, the target prop horse is controlled to move together with the target human body riding on the horse through the movement and shaking of the hip joint or shoulder joint, etc.), but also can liberate the upper limbs and other parts that need to undertake the game's hitting action tasks, and can also give real-life games more fitness attributes. In addition, since the depth (forward and backward movement direction) range of the hips / knees / shoulders and other parts when moving and shaking is very small, it is easier to identify their movement amplitude based on ordinary camera devices and plane images.

[0070] However, due to the characteristics of human body structure and the limitations of existing key point recognition technology for human joints, even if the above-mentioned joints move to a large extent, the position change distance of the identifiable key points corresponding to the joints on the human body plane image is generally not large, which makes it difficult to accurately identify the movement amplitude of the joints that conforms to the user's body perception. When the joints move with the same amplitude, the position change distance of the intersection of the scan line passing through the joints and the target human body boundary contour line (i.e., the edge point of the joint) on the human body plane image is generally larger. Among them, the position change distance of the intersection of the scan line that is basically consistent with the movement trajectory of the joints and passes through the joints and the target human body boundary contour line on the human body plane image is relatively larger. Therefore, the movement amplitude of the joints that is consistent with the user's body perception can be more accurately identified.

[0071] At the same time, because the existing technology has certain errors in obtaining the position of the same identifiable key points in different frames of human body images, when the position change distance of the position feature points is small, such errors will cause greater calculation interference, while when the position change distance of the position feature points is large, the interference caused by such errors can be basically ignored.

[0072] The present invention uses the intersection of the scanning line L and the boundary contour line of the target human body to obtain the position feature points P1 and P2. Compared with directly identifying the joint feature points of the human body joints, it is easier to obtain the movement amplitude of the joints that matches the user's body sensation, and the error caused when obtaining the same identifiable key point position of different frames of human body images is also small, which can ensure the user's good body sensation when playing games.

[0073] Furthermore, based on the relative position parameters of the position feature point P1 of the current frame human body image and the position feature point P2 of the comparison frame human body image, the current movement amplitude S is calculated. If the movement amplitude S between the comparison frame human body image and the current frame human body image satisfies a preset condition, such as the movement amplitude exceeds a certain threshold, it means that the amplitude of the movement of a certain part of the human body exceeds the threshold. Based on this, it can be judged that the user is sending an action signal to drive the prop to move, thereby triggering the prop movement instruction and changing the visual relative position of the prop and / or the target human body in the human body image in the target picture based on the same rule.

[0074] It should be noted that the position feature point can be one or more. For example, a position feature point corresponding to the left knee or the right knee can be taken, or the feature points corresponding to the left knee and the right knee can be taken at the same time. In addition, the boundary contour line of the target human body can be the outer contour line and / or the inner contour line of the human body, which can be selected according to the specific calculation accuracy and calculation needs.

[0075] It should be noted that there are many ways to change the visual relative position of the target props and / or the target human body in the human body image in the target screen, such as the game screen. For example, the position of the target props, such as the human body-carrying props and the target human body it carries, in the game screen can be changed as a whole to obtain the visual effect of the overall movement of the human body-carrying props and the target human body it carries. The visual effect of the overall movement of the human body-carrying props and the target human body it carries can also be obtained by changing other contents in the game screen, such as the background scenery and the relative position of the human body-carrying props and the target human body it carries. In addition, only the visual relative position of the target human body or the target props in the target screen can be changed. For example, in a game without human body-carrying props, the user triggers his own image (i.e., the target human body) through specific actions to complete the movement required by the game plot, etc. The present invention is not limited to this.

[0076] Preferably, the comparison frame human body image is a frame human body image among K frames human body image before the current frame human body image, or the comparison frame human body image is a frame human body image among several frames human body image within a time length K before the current frame human body image.

[0077] By adopting the above technical solution, the following technical effects are achieved:

[0078] Based on the relative position parameters of the position feature point P1 of the current frame human body image and the position feature point P2 on the comparison frame human body image in the previous K frames or the historical human body images within the previous K time, a movement amplitude S is generated. If the movement amplitude S between the comparison frame human body image and the current frame human body image satisfies the preset conditions, such as the movement amplitude exceeds a certain threshold, it means that the speed and amplitude of the movement of a certain part of the human body exceed the threshold, and such a part movement will not occur under normal conditions. This can avoid interference from other irrelevant movements of the human body, achieve user perception of stable movement expectations and real-time movement responses, and reduce the user's sense of restraint when playing games. It should be noted that the current frame / comparison frame human body image can be one frame or multiple frames. For example, the average value of the corresponding position information of the position feature points of the adjacent frames corresponding to the current frame / comparison frame human body image is used as the basis for subsequent calculations to reduce possible information acquisition errors of the camera device.

[0079] like Figure 2 As shown, in this embodiment, when calculating the current movement amplitude S, the position feature point P1 and the position feature point P2 are projected onto the preset trajectory line M, and the movement amplitude S is generated based on the segment of the preset trajectory line M between the projection point P1' of the position feature point P1 and the projection point P2' of the position feature point P2.

[0080] By adopting the above technical solution, the following technical effects are achieved:

[0081] For the convenience of users, the preset trajectory line M is usually a reference movement direction or reference movement trajectory of the target part to prompt the user in the game, such as a preset reference human body movement. Since the actual human body movement is unlikely to be completely consistent with the movement direction of the preset reference human body movement, the corresponding projection point P1' and projection point P2' are obtained by projecting the position feature point P1 and the position feature point P2 onto the preset trajectory line M, and then the movement amplitude S is generated based on the segment of the preset trajectory line M between P1' and P2'. The movement amplitude of the target part can be judged more concisely and the user's sense of restraint during the game can be further reduced.

[0082] Furthermore, the scanning line L and the preset trajectory line M are parallel.

[0083] By adopting the above technical solution, the following technical effects are achieved:

[0084] The target human body will move the target part in the direction of the preset trajectory line M according to the prompt, and the scanning line that is basically consistent with the target part movement direction and passes through the target part, and the intersection of the target body boundary contour line and the target body boundary contour line has a relatively larger position change distance based on the human body plane image during the movement process, so the target part movement amplitude that conforms to the user's body feeling can be more accurately identified. In addition, since the position feature point is obtained based on the intersection of the scanning line that is basically consistent with the target part movement direction and the target body boundary contour line, the interference caused by the position calculation error of the identifiable key points of the human body images in different frames in the target part movement direction can be shielded, and the calculation accuracy of the movement amplitude S is further improved. It should be noted that the scanning line L can also be set to be parallel to the line between the identifiable feature points of the target part of the current frame and the corresponding identifiable feature points of the target part of the comparison frame, which is essentially the same.

[0085] like Figure 2 As shown, in this embodiment, the movement amplitude S includes a movement direction T and a movement distance D.

[0086] The moving direction T represents the moving direction of the target part, and can be generated in a variety of ways. For example, when the angle between the "connecting line or the direction vector of the connecting line" and the "horizontal line or other reference line" of the position feature point P2 in the comparison frame human body image and the position feature point P1 in the current frame human body image is within a preset range, it is positive, otherwise it is reverse; for example, when the projection point P2' of the position feature point P2 of the comparison frame human body image on the preset trajectory line M is closer to the starting point of the preset trajectory line M (such as the intersection of the preset trajectory line M and the boundary line of the target image) than the projection point P1' of the position feature point P1 of the current frame human body image on the preset trajectory line M, then the moving direction T1 is positive, otherwise it is reverse, etc.;

[0087] The moving distance D represents the moving distance of the target part, and is generated based on the relative position parameters of the position feature point P2 of the comparison frame human body image and the position feature point P1 of the current frame human body image. It should be noted that the relative position parameters can be obtained in various ways, such as based on the relative position distance between the position feature points P2 and P1 on the human body plane image or the relative position distance between the projection points P2' and P1' of the position feature points P2 and P1 on the preset trajectory line M.

[0088] In this embodiment, the movement amplitude S satisfies the preset condition when: the movement direction T is within a preset movement direction range (for example, the movement direction T is positive), and the movement distance D is greater than the preset movement distance D′.

[0089] By adopting the above technical solution, the moving direction and moving distance of the target part can be identified at the same time, and the judgment of the movement of the target part can be more accurate.

[0090] Preferably, the moving distance D is generated based on the ratio of the relative position parameter between the position feature point P1 and the position feature point P2 to the size of a specific part of the human body. The advantage of this is that, since the body shape characteristics of actual human bodies vary greatly in practice, and the distances of actual human bodies from the camera device are also different, this method can convert the moving distance D of the target part of actual human bodies of different body shapes / different distances from the camera device into the same measurement system, eliminating the change in the moving distance D caused by different environments, so that there is no need to set a preset moving distance D' for each environment.

[0091] Further optionally, the size of the specific human body part is a relative position parameter between the position feature point P1 and the human body support point U1 on the current frame human body image, and the human body support point U1 is a identifiable feature point corresponding to the human body support part on the target human body.

[0092] It should be noted that the human body support point U1 is an identifiable feature point corresponding to a fixed support part of the human body that supports the human body and basically does not move when the target part moves, such as the foot feature point when the knee and hip move and shake, such as the hip feature point when the shoulder moves and shakes, etc., which can be selected according to actual needs.

[0093] The advantage of doing so is that, according to the structural characteristics of the human body, when the joints such as the knees and hips move and shake, there is a direct correlation between the upper limit of the movement amplitude and the upper limit of the range of change of the angle between the position feature point corresponding to the joint and the human support point corresponding to the fixed support part on the target human body. Therefore, based on the relative position parameters between the position feature point P1 and the human support point U1, the corresponding size of the specific part of the human body is obtained, and then the corresponding moving distance D is calculated, which can eliminate the interference caused by body shape differences in the calculation of the moving distance D to the greatest extent. It should be noted that the size of the specific part of the human body can also be the length of the segment of the scanning line L in the specific part of the human body. In addition, the size of the specific part of the human body can also be obtained based on the relative position parameters between the corresponding position feature points and the human support points on the human body images of the comparison frame or other frames, which is essentially the same.

[0094] like Figure 3 As shown, in another implementation of the present embodiment, the moving distance D is generated based on the angle C between the line O1 and the line O2, the line O1 refers to the line connecting the position feature point P1 and the human body support point U1 on the current frame human body image, the line O2 refers to the line connecting the position feature point P2 and the human body support point U2 on the comparison frame human body image, and the human body support point U1 and the human body support point U2 are both identifiable feature points corresponding to the human body support parts on the target human body.

[0095] According to the structural characteristics of the human body, when joints such as the knees and hips move and shake, there is a direct corresponding relationship between the movement distance and the change in the angle between the position feature points corresponding to the joints and the human support points corresponding to the fixed support parts on the target human body. Therefore, obtaining the movement distance D based on the relative position parameters between the position feature points P1 and P2 and obtaining the movement distance D based on the angle C between the connection line O1 and the connection line O2 are essentially the same.

[0096] It should be noted that the moving distance D is obtained based on the angle C between the line O1 and the line O2. In practice, the angle C can correspond to different distance values ​​within different angle ranges. For example, the angle C at 0°-10° corresponds to a distance value of D1, the angle C at 11°-20° corresponds to a distance value of D2, the angle C at 21°-30° corresponds to a distance value of D3, and so on.

[0097] It should be noted that the identifiable feature points are feature points corresponding to human body parts that can be obtained through image recognition, including feature points corresponding to human body parts that can be obtained through image recognition or feature points converted based on the relative position relationship between the feature points corresponding to human body parts and “other feature points corresponding to human body parts” / “points or lines on the human body image” / “points or lines on the target screen”, etc.;

[0098] Obtaining corresponding feature points of human body parts through image recognition is a commonly used technology in this field. For example, corresponding feature points of human body parts in human images can be obtained through human body recognition software implemented based on deep learning or other methods, or corresponding areas of the human body can be obtained through human body recognition software implemented based on deep learning or other methods, and then corresponding feature points of human body parts can be obtained through image analysis. Corresponding feature points of human body parts can also be obtained by setting marking lines or markers, such as setting marking points / marking lines on the display screen, prompting the user to make corresponding feature points of human body parts such as the vertex of the head / vertical midline of the human body coincide with corresponding marking points / marking lines based on their own perception to obtain corresponding feature points of human body parts, or marking specific markers on specific parts of the human body and then obtaining corresponding feature points of human body parts through image analysis, etc.

[0099] Preferably, the method further includes executing the following steps before triggering the prop movement instruction:

[0100] Step 1, mark the current frame human body image as the comparison frame human body image;

[0101] Step 2, read the next frame of human body image as the new current frame of human body image;

[0102] Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image;

[0103] Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame;

[0104] If the moving distance D>preset threshold a, and the moving direction T is positive, then mark the current frame of human body image as the same direction frame;

[0105] If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame;

[0106] Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

[0107] By adopting the above technical solution, the following technical effects are achieved:

[0108] When the movement amplitude S of the target part meets the preset conditions and moves to the end based on the same direction (that is, a complete target part movement action is completed), the prop movement instruction is triggered, which can achieve better user physical perception and picture effects. At the same time, after the hip, knee, shoulder and other parts complete the movement in a certain direction, the corresponding part of the human body will instinctively reset back to the initial relatively comfortable position. At this time, it is easy to misidentify the reset action as a movement in the opposite direction, thereby triggering the wrong prop movement instruction. After the prop movement instruction is triggered, the game generally sets the action recognition sleep time to ensure that the picture movement effect is completed first. Therefore, when the target part moves to the end based on the same direction, the prop movement instruction is triggered again, which can make the reset action time and the action recognition sleep time overlap to avoid the above-mentioned action misrecognition. It should be noted that in actual operation, more preset conditions can be added to trigger the prop movement instruction in addition to the preset conditions when the movement amplitude S of the target part meets the preset conditions to avoid the above-mentioned misrecognition. For example, the prop movement instruction will be triggered only when the position characteristics or shape characteristics of the target part meet specific conditions.

[0109] It should be further explained that, since the prior art has certain deviations when obtaining identifiable feature points of the human body in different frames of human body images, when the moving distance D ≤ the preset threshold a, the current frame of the human body image is marked as a still frame. Furthermore, if n still frames and / or reverse direction frames appear continuously, the prop movement instruction is triggered, and n is the threshold. This can effectively avoid the interference caused by the position recognition error of the identifiable feature points of the human body.

[0110] Example 2

[0111] like Figure 4 As shown, a target prop movement triggering system based on human body image comparison includes a first acquisition module, a second acquisition module, a calculation module, a judgment module and an execution module;

[0112] The first acquisition module is used to acquire a position feature point P1 of a human body image of a current frame; the position feature point P1 is obtained based on an intersection of a scanning line L and a boundary contour line of a target human body in the human body image of the current frame;

[0113] The second acquisition module is used to acquire a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scanning line L and the boundary contour line of the target human body in the comparison frame human body image;

[0114] It should be noted that the first acquisition module and the second acquisition module may be the same module or different modules;

[0115] The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image;

[0116] The calculation module is used to calculate the movement amplitude S of the human body image of the current frame based on the position feature point P1 and the position feature point P2;

[0117] The judging module is used to judge whether the movement amplitude S of the human body image in the current frame meets the preset conditions, and if so, trigger the prop movement instruction;

[0118] The execution module is used to change the visual relative position of the target prop and / or the target human body in the target picture based on the same rule after triggering the prop movement instruction.

[0119] In this embodiment, the comparison frame human body image is a frame human body image among the K frames human body image before the current frame human body image, or the comparison frame human body image is a frame human body image among several frames human body image within a time length K before the current frame human body image.

[0120] like Figure 2 As shown, in this embodiment, the movement amplitude S includes a movement direction T and a movement distance D.

[0121] Among them, the moving direction T includes a forward direction and a reverse direction. When calculating the moving direction T, the position feature point P1 and the position feature point P2 are projected onto the preset trajectory line M to generate the projection point P1' of the position feature point P1 and the projection point P2' of the position feature point P2. If the projection point P1' is closer to the starting point of the preset trajectory line M than the projection point P2', the moving direction T is forward, otherwise the moving direction T is reverse.

[0122] The moving distance D is generated by the ratio of the length of the segment of the preset trajectory line M between the projection point P1 ′ of the position feature point P1 and the projection point P2 ′ of the position feature point P2 to the size of the specific part of the human body.

[0123] The moving distance D may also be generated based on the ratio of the relative position parameter between the position feature point P1 and the position feature point P2 to the size of a specific part of the human body.

[0124] It should be noted that the size of the specific human body part is the relative position parameter between the position feature point P1 and the human body support point U1 on the current frame human body image, and the human body support point U1 is a recognizable feature point corresponding to the human body support part on the target human body.

[0125] like Figure 3As shown, the moving distance D can also be generated based on the angle C between the line O1 and the line O2, wherein the line O1 refers to the line connecting the position feature point P1 and the human body support point U1 on the current frame human body image, and the line O2 refers to the line connecting the position feature point P2 and the human body support point U2 on the comparison frame human body image, and the human body support point U1 and the human body support point U2 are both identifiable feature points corresponding to the human body support parts on the target human body.

[0126] In this embodiment, the movement amplitude S includes a movement direction T and a movement distance D;

[0127] The judgment module performs the following steps before triggering the prop movement instruction:

[0128] Step 1, mark the current frame human body image as the comparison frame human body image;

[0129] Step 2, read the next frame of human body image as the new current frame of human body image;

[0130] Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image;

[0131] Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame;

[0132] If the moving distance D>preset threshold a, and the moving direction T is positive, then mark the current frame of human body image as the same direction frame;

[0133] If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame;

[0134] Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

[0135] In order to briefly explain the implementation method and principle of Example 2, some explanations mentioned in Example 1 are omitted. The technical principles of Example 1 and Example 2 are the same. If there are any unclear parts in Example 2, please refer to Example 1.

[0136] The above is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment according to the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention. The protection scope of the present invention shall be subject to the protection scope of the claims and their equivalents.

Claims

1. A target movement triggering method based on human body image comparison, characterized in that: The following steps are involved: Acquire a position feature point P1 of the human body image of the current frame; the position feature point P1 is obtained based on the intersection of the scan line L and the boundary contour line of the target human body in the human body image of the current frame; Acquire a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scan line L and the boundary contour line of the target human body in the comparison frame human body image; The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image; Based on the position feature point P1 and the position feature point P2, the movement amplitude S of the human body image in the current frame is calculated; Determine whether the movement amplitude S of the human body image in the current frame meets the preset conditions. If so, trigger the prop movement instruction; After the prop movement command is triggered, the visual relative positions of the target prop and the target human body in the target screen are changed based on the same rule; When calculating the current movement amplitude S, the position feature points P1 and P2 are projected onto the preset trajectory line M, and the movement amplitude S is generated based on the segment of the preset trajectory line M between the projection point P1' of the position feature point P1 and the projection point P2' of the position feature point P2.

2. A target movement triggering method based on human body image comparison according to claim 1, characterized in that: The comparison frame human body image is a human body image among K human body image frames before the current human body image frame, or the comparison frame human body image is a human body image among several human body image frames within a time length K before the current human body image frame.

3. A target movement triggering method based on human body image comparison according to claim 1, characterized in that: The scanning line L is parallel to the preset trajectory line M.

4. A target movement triggering method based on human body image comparison according to claim 1 or 2, characterized in that: The movement amplitude S includes a movement direction T and a movement distance D.

5. A target movement triggering method based on human body image comparison according to claim 4, characterized in that: The moving distance D is generated based on the ratio of the relative position parameter between the position feature point P1 and the position feature point P2 to the size of the human body part.

6. A method for triggering target movement based on human body image comparison according to claim 5, characterized in that: The human body part size is a relative position parameter between the position feature point P1 and the human body support point U1 on the current frame human body image, and the human body support point U1 is a recognizable feature point corresponding to the human body support part on the target human body.

7. A method for triggering target movement based on human body image comparison according to claim 6, characterized in that: The moving distance D is generated based on the angle C between the line O1 and the line O2, the line O1 refers to the line connecting the position feature point P1 and the human body support point U1 on the current frame human body image, and the line O2 refers to the line connecting the position feature point P2 and the human body support point U2 on the comparison frame human body image, and the human body support point U1 and the human body support point U2 are both identifiable feature points corresponding to the human body support parts on the target human body.

8. A method for triggering target movement based on human body image comparison according to claim 7, characterized in that: It also includes executing the following steps before triggering the prop movement command: Step 1, mark the current frame human body image as the comparison frame human body image; Step 2, read the next frame of human body image as the new current frame of human body image; Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image; Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame; If the moving distance D> the preset threshold a, and the moving direction T is positive, then mark the current frame of the human body image as a same-direction frame; If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame; Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

9. A target movement triggering system based on human body image comparison, characterized in that: It includes a first acquisition module, a second acquisition module, a calculation module, a judgment module and an execution module; The first acquisition module is used to acquire a position feature point P1 of a human body image of a current frame; the position feature point P1 is obtained based on an intersection of a scanning line L and a boundary contour line of a target human body in the human body image of the current frame; The second acquisition module is used to acquire a position feature point P2 of the comparison frame human body image; the position feature point P2 is obtained based on the intersection of the scanning line L and the boundary contour line of the target human body in the comparison frame human body image; The scanning line L is a scanning line passing through a target part of a target human body in a current frame human body image or a comparison frame human body image; The calculation module is used to calculate the movement amplitude S of the human body image of the current frame based on the position feature point P1 and the position feature point P2; When calculating the current movement amplitude S, the position feature point P1 and the position feature point P2 are projected onto the preset trajectory line M, and the movement amplitude S is generated based on the segment of the preset trajectory line M between the projection point P1' of the position feature point P1 and the projection point P2' of the position feature point P2 on the preset trajectory line M; The judging module is used to judge whether the movement amplitude S of the human body image in the current frame meets the preset conditions, and if so, trigger the prop movement instruction; The execution module is used to change the visual relative positions of the target prop and the target human body in the target picture based on the same rule after triggering the prop movement instruction.

10. A target movement triggering system based on human body image comparison according to claim 9, characterized in that: The comparison frame human body image is a human body image among K human body image frames before the current human body image frame, or the comparison frame human body image is a human body image among several human body image frames within a time length K before the current human body image frame.

11. A target movement triggering system based on human body image comparison according to claim 9, characterized in that: The movement amplitude S includes a movement direction T and a movement distance D; The judgment module performs the following steps before triggering the prop movement instruction: Step 1, mark the current frame human body image as the comparison frame human body image; Step 2, read the next frame of human body image as the new current frame of human body image; Step 3, based on the current frame human body image and the comparison frame human body image, calculate the moving direction T and moving distance D of the current frame human body image; Step 4, if the moving distance D is less than or equal to the preset threshold a, the current frame of the human body image is marked as a still frame; If the moving distance D> the preset threshold a, and the moving direction T is positive, then mark the current frame of the human body image as a same-direction frame; If the moving distance D>preset threshold a, and the moving direction T is in the reverse direction, then mark the current frame of human body image as a reverse direction frame; Step 5. Determine whether n still frames and / or reverse direction frames appear continuously. If so, trigger the prop movement command. If not, go to Step 1.

Images