Frame insertion method and device for face recognition tracking display

By predicting the position of the face frame and using linear interpolation method, the problem of low fluency of the face recognition interface is solved, and the interface fluency and prediction accuracy are improved under low frame rate, reducing costs and enhancing the system robustness.

CN119049099APending Publication Date: 2024-11-29XIAMEN STAR SMART TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410918404.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Due to the large amount of face detection in facial recognition, the frame rate of the output face frame is reduced, resulting in the problem of low interface fluency. Although existing methods such as multi-threading or hardware upgrades can improve, they are costly.

Method used

By predicting the face frame position between the current frame and the future frame based on the face information of the past frame and the current frame, and using linear interpolation method to display the predicted frame information, including calculating the movement speed and position of the face frame, limiting the speed in the real world The face swings within the range to prevent the face frame from deviating.

Benefits of technology

It effectively improves the fluency of the face recognition interface at low frame rates, predicts the position of the face frame accurately, reduces costs, improves the robustness of the system, and avoids the problem of the face frame deviating when suddenly reversed.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119049099A_ABST
    Figure CN119049099A_ABST
Patent Text Reader

Abstract

The invention discloses a frame insertion method and device for face recognition tracking display, and relates to the technical field of face recognition. The method comprises the following steps: step 1, according to a face detection result, storing and continuously updating position information of a face frame of a previous frame and face detection time; step 2, estimating face frame position information of a prediction frame of at least one specified time between the current frame and the future frame according to the face frame position information of the previous frame, the face detection time and the face frame position information of the current frame; and step 3, displaying the face frame according to the face frame position information at specified time. According to the frame insertion method and device for face recognition tracking display provided by the invention, the face frame position between the current frame and the future frame is predicted according to the face information of the past frame and the current frame, and the predicted frame information is displayed, so that the problem of low fluency of a face recognition interface under the condition of low frame rate can be well improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of face recognition technology, and in particular to a frame insertion method and device for face recognition tracking display. Background Art

[0002] In face recognition, since face detection involves a large amount of computation and the entire process is time-consuming, the frame rate of the output face frame is reduced, resulting in the face frame output to the device interface not being smooth enough for the user to perceive.

[0003] Existing methods, such as using multi-threading or multi-processing to increase the frame rate, can reduce the lag problem to a certain extent; or upgrading hardware equipment, adding GPUs, speeding up face detection, and increasing frame rates, although the effect is good, the cost is high. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a frame insertion method and device for face recognition tracking display, which predicts the face frame position between the current frame and the future frame based on the face information of the past frame and the current frame, and displays the predicted frame information, which can better improve the problem of low fluency of the face recognition interface under low frame rate conditions.

[0005] In a first aspect, the present invention provides a frame insertion method for face recognition tracking display, comprising:

[0006] Step 1: According to the result of face detection, the face frame position information and face detection time of the previous frame are stored and continuously updated;

[0007] Step 2: Estimate the face frame position information of a predicted frame at least one specified time between the current frame and the future frame according to the face frame position information of the previous frame and the face detection time and the face frame position information of the current frame;

[0008] Step 3: Display the face frame according to the face frame position information at the specified time.

[0009] Furthermore, the step 2 specifically includes:

[0010] Step 21, calculate the moving speed of the face frame in the horizontal coordinate direction, then regard the moving trajectory of the face frame between two frames as uniform linear motion, and calculate the moving speed v of the face frame in the horizontal coordinate direction;

[0011] Step 22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v;

[0012] Step 23: Calculate the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

[0013] Furthermore, in step 22, a speed limit area is predicted based on the lateral swing range of the real face, and the moving speed v is limited within the speed limit area.

[0014] Furthermore, predicting the speed limit area according to the lateral swing range of the real face and limiting the moving speed v in the speed limit area specifically includes:

[0015] Assume that the initial position of the static face frame is x0, the width of the face frame is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w.

[0016] The left extreme position of face swing to the right extreme position of face swing is taken as the real face lateral swing range, and the area outside the real face lateral swing range is set as the speed limit area, and the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1;

[0017] When the face frame is in a stationary state, the initial position x0 is reset.

[0018] Furthermore, the time difference between the predicted frame and the current frame is smaller than the time difference between the current frame and the past frame.

[0019] In a second aspect, the present invention provides a frame insertion device for face recognition tracking display, comprising:

[0020] A storage and update module is used to store the face frame position information and face detection time of the previous frame and continuously update them according to the face detection result;

[0021] A prediction frame module, used to estimate the face frame position information of a prediction frame of at least a specified time between the current frame and the future frame according to the face frame position information and face detection time of the previous frame and the face frame position information of the current frame;

[0022] The display module is used to display the face frame according to the face frame position information at a specified time.

[0023] Furthermore, the prediction frame module is specifically used to perform the following steps:

[0024] Step 21, calculate the moving speed of the face frame in the horizontal coordinate direction, then regard the moving trajectory of the face frame between two frames as uniform linear motion, and calculate the moving speed v of the face frame in the horizontal coordinate direction;

[0025] Step 22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v;

[0026] Step 23: Calculate the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

[0027] Furthermore, in step 22, a speed limit area is predicted based on the lateral swing range of the real face, and the moving speed v is limited within the speed limit area.

[0028] Furthermore, predicting the speed limit area according to the lateral swing range of the real face and limiting the moving speed v in the speed limit area specifically includes:

[0029] Assume that the initial position of the static face frame is x0, the width of the face frame is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w.

[0030] The area from the left extreme position of face swing to the right extreme position of face swing is set as the speed limit area; the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1;

[0031] When the face frame is in a stationary state, the initial position x0 is reset.

[0032] Furthermore, the time difference between the predicted frame and the current frame is smaller than the time difference between the current frame and the past frame.

[0033] The technical solution provided in the embodiments of the present invention has at least the following technical effects:

[0034] The method of predicting the position of the face frame and interpolating frames can effectively improve the problem of low smoothness of the face recognition interface under low frame rate conditions. The linear interpolation method is used to predict the face frame. Under normal circumstances, the position of the face frame can be well predicted.

[0035] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The present invention will be further described below in conjunction with embodiments with reference to the accompanying drawings.

[0037] Figure 1 It is a schematic diagram of the principle of an embodiment of the present invention;

[0038] Figure 2 is an overall flow chart of the method in Embodiment 1 of the present invention;

[0039] Figure 3Schematic diagram of the horizontal coordinate x of the face frame of the estimated prediction frame in the first embodiment of the present invention;

[0040] Figure 4 Schematic diagram of the vertical coordinate x of the face frame of the estimated prediction frame in the first embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram of a face frame deviating from a face when a face suddenly moves in the opposite direction during the face swinging process in the first embodiment of the present invention;

[0042] Figure 6 Schematic diagram of the range of extreme positions of face swing in Embodiment 1 of the present invention;

[0043] Figure 7 It is a schematic diagram of the structure of the device in the second embodiment of the present invention. DETAILED DESCRIPTION

[0044] The embodiment of the present invention provides a frame insertion method and device for face recognition tracking display, predicts the face frame position between the current frame and the future frame based on the face information of the past frame and the current frame, and displays the predicted frame information, which can effectively improve the problem of low fluency of the face recognition interface under low frame rate conditions.

[0045] The technical solution in the embodiment of the present invention has the following general idea:

[0046] A low-cost, effective and robust interpolation method for face recognition tracking display is proposed. The face frame position at a certain time between the current frame and the future frame is predicted through the face frame information of the past frame and the current frame, and the predicted frame information is displayed. Figure 1 shown.

[0047] The relevant steps are as follows:

[0048] 1. Store face information in past frames.

[0049] 2. Estimate the face frame position information of the predicted frame:

[0050] a) Estimate the moving speed of the face frame in the horizontal coordinate direction;

[0051] b) estimating the horizontal coordinate position of the face frame of the prediction frame;

[0052] c) estimating the vertical coordinate position of the face frame of the prediction frame;

[0053] d) Return the estimated face frame position information of the predicted frame.

[0054] 3. Estimate the speed limit area based on the real face lateral swing trajectory and limit the speed in this area:

[0055] a) Estimate the speed limit area based on the lateral swing trajectory of the real face;

[0056] b) Testing appropriate speed limit thresholds;

[0057] c) Resetting the initial position when the face is still.

[0058] Embodiment 1

[0059] This embodiment provides a frame insertion method for face recognition tracking display, such as Figure 2 As shown, the following steps may be included:

[0060] S1. According to the result of face detection, the face frame position information and face detection time of the previous frame are stored and continuously updated;

[0061] S2, estimating face frame position information of a predicted frame at a specified time between the current frame and the future frame according to the face frame position information of the previous frame and the face detection time and the face frame position information of the current frame; wherein the number of predicted frames may be one or more than one;

[0062] S3. Displaying the face frame according to the face frame position information at a specified time.

[0063] The method of predicting the position of the face frame and interpolating frames can effectively improve the problem of low smoothness of the face recognition interface under low frame rate conditions. The linear interpolation method is used to predict the face frame. Under normal circumstances, the position of the face frame can be well predicted.

[0064] Preferably, the S2 specifically includes:

[0065] S21, calculating the moving speed of the face frame in the horizontal coordinate direction, and then considering the moving trajectory of the face frame between two frames as uniform linear motion, and calculating the moving speed v of the face frame in the horizontal coordinate direction;

[0066] S22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v;

[0067] S23, calculating the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

[0068] Preferably, in S22, the speed limit area is predicted according to the lateral swing range of the real face, and the moving speed v is limited in the speed limit area, specifically including:

[0069] Assume that the initial position of the static face frame is x0, the width of the face frame is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w.

[0070] The left extreme position of face swing to the right extreme position of face swing is taken as the real face lateral swing range, and the area outside the real face lateral swing range is set as the speed limit area, and the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1;

[0071] When the face frame is in a stationary state, the initial position x0 is reset.

[0072] The range where speed limitation is required is estimated according to the maximum swing amplitude of the real face, and the speed limitation is applied to it. This method improves the problem that the face frame deviates from the real face when the face suddenly moves in the opposite direction, and improves the accuracy of predicting the face frame position.

[0073] In a specific embodiment, taking inserting a predicted frame between the current frame and the future frame as an example, the implementation process is as follows:

[0074] 1. Store face information in past frames:

[0075] According to the result of face detection, the face frame position information of the previous frame and the time of face detection are continuously updated and stored.

[0076] 2. Calculate the face frame position information of the predicted frame:

[0077] The face frame position information of the previous frame (ie, the past frame) and the face frame position information of the current frame are used to estimate the face frame position information at a random time between the current frame and the future frame.

[0078] 1. Estimate the moving speed of the face frame in the horizontal coordinate direction:

[0079] Assume that the moving speed of the face frame in the horizontal direction is v. In face recognition, the time interval between two adjacent frames is very short, about 40ms. In a very short time, the movement of the face frame in the horizontal direction can be approximately regarded as uniform linear motion. Therefore, the moving speed v of the face frame in the horizontal direction can be estimated by the relationship v=S / t.

[0080] a) Calculate the moving distance of the face frame in the horizontal coordinate direction of two adjacent frames:

[0081] Assume that the horizontal coordinates of the upper left corner of the face frame of the previous frame and the current frame are x1 and x2 respectively, and the moving distance of the face frame of two adjacent frames in the horizontal coordinate direction is S. S is calculated by x1-x2, and the sign before the calculation result indicates the direction of face movement (positive sign means movement to the right, negative sign means movement to the left).

[0082] The absolute value of the distance S should be set to be less than the maximum horizontal movement distance of a normal face in the interval between two adjacent frames (this value needs to be obtained based on actual testing), otherwise the face frame offset problem may occur when switching between different faces at a long distance. By filtering the large changes in the horizontal coordinate position of the face frame between the past frame and the current frame, the face offset problem that occurs when different faces are switched at a long distance is solved, and the robustness of the system is improved.

[0083] b) Calculate the moving time between two adjacent frames:

[0084] Assume that the face detection time of the past frame and the current frame is T1 and T2 respectively, and the moving time between two adjacent frames is △T1. Take the difference between T1 and T2 to calculate △T1.

[0085] 2. Estimate the horizontal coordinate position of the face frame of the predicted frame:

[0086] a) Assume that the time for obtaining the predicted frame and calling face detection is T3, the horizontal coordinate position of the predicted frame face frame is x, and the time interval between the predicted frame and the current frame is △T2. △T2 is obtained by subtracting T3 and T2. Note: △T2 needs to be set smaller than △T1, otherwise the face frame will stay for a long time when the face suddenly disappears from the screen.

[0087] b) Estimate x by x2+v×△T2. Figure 3 shown.

[0088] 3. Estimate the vertical coordinate position of the face frame of the predicted frame:

[0089] a) Because the time interval between two adjacent frames is short enough, the moving track of the face frame between the two frames can be approximately regarded as a straight line within the time interval. Therefore, the present invention uses a linear interpolation method to estimate the vertical coordinate position of the face frame of the prediction frame.

[0090] b) Assume that the ordinates of the upper left corner of the face frame in the previous frame and the current frame are y1 and y2 respectively, and the slope of the face frame moving track between the two frames is k. Then k = (y2-y1) / (x2-x1). Calculate the linear equation of the moving track.

[0091] c) Assume that the vertical coordinate of the predicted face frame is y. Use the horizontal coordinate position x of the predicted face frame calculated previously and the linear equation of the moving trajectory obtained above to estimate y, as follows: Figure 4 shown.

[0092] 4. Return the estimated face frame position information of the predicted frame.

[0093] 3. Estimate the speed limit area based on the horizontal swing trajectory of the real face, and limit the speed in this area:

[0094] Because the prediction calculation method for the predicted frame face frame position is linear interpolation, this method will have the problem of the predicted frame face frame deviating from the real face when the speed is fast and the reverse movement is sudden, resulting in poor performance of the face recognition interface in this case, such as Figure 5 As shown in Figure 2, a solution is proposed to estimate the speed limit area according to the horizontal swing trajectory of the real face and slow down the speed in this area.

[0095] 1. Estimating the speed limit area based on the horizontal swing trajectory of the real face:

[0096] In the device display interface, the distance between the face and the device will affect the extreme position range of the left and right sides of the real face swing. In face recognition, the width of the face frame will change with the distance between the face and the device. Therefore, the speed limit area can be measured based on the width of the face frame.

[0097] Assume that the initial position of the static state is x0, the face frame width is w, and the width threshold is n (indicates how many face frame widths are required from the initial position to the left or right boundary). By testing the left and right extreme positions of the face swing multiple times, such as Figure 6 As shown, to estimate n.

[0098] a) The left extreme position is x0-n×w.

[0099] b) The right extreme position is x0+n×w.

[0100] 2. Test the appropriate speed limit threshold:

[0101] a) The fundamental reason for the face frame offset is that the speed of the face frame in the predicted frame is estimated based on the information of the first two adjacent frames, which has a certain delay. If the real face suddenly moves in the opposite direction, the face frame in the predicted frame will continue to move a short distance in the original direction, and the reverse movement of the real face will also increase the distance between the face frame and the face. Therefore, limiting the speed of the part outside the previously measured motion range can improve the problem of face frame offset to a certain extent.

[0102] b) Assuming the speed limit threshold is m, the speed after the limit is v×m. Through multiple tests on the device, a suitable m is selected (for example, 1 / 10, which can be adjusted according to the actual face tracking situation) while minimizing the deviation of the face frame and not affecting the face frame and face in the normal range.

[0103] 3. Reset the initial position when the face is still:

[0104] a) Set the horizontal coordinate threshold a and the vertical coordinate threshold b. In this embodiment, the judgment standard for the static state is that the absolute value of the difference between the horizontal coordinates of the face frame between two adjacent frames must be less than a, and the same applies to the vertical coordinate. The values ​​of a and b need to be determined through multiple tests based on the accuracy of the static state judgment and the reaction speed of the initial position reset. Otherwise, there may be a problem that x0 is reset during the movement or x0 cannot be reset in time after being stationary.

[0105] b) After the above conditions determine that the state is static, the position at the static state is set as the initial position.

[0106] The range where speed limitation is required is estimated according to the maximum swing amplitude of the real face, and the speed limitation is applied to it. This method improves the problem that the face frame deviates from the real face when the face suddenly moves in the opposite direction, and improves the accuracy of predicting the face frame position.

[0107] Based on the same inventive concept, the present application also provides a device corresponding to the method in Example 1, see Example 2 for details.

[0108] Embodiment 2

[0109] In this embodiment, a frame insertion device for face recognition tracking display is provided. Figure 7 As shown, including:

[0110] A storage and update module is used to store the face frame position information and face detection time of the previous frame and continuously update them according to the face detection result;

[0111] A prediction frame module, used to estimate the face frame position information of a prediction frame of at least a specified time between the current frame and the future frame according to the face frame position information and face detection time of the previous frame and the face frame position information of the current frame;

[0112] The display module is used to display the face frame according to the face frame position information at a specified time.

[0113] Preferably, the prediction frame module is specifically used to perform the following steps:

[0114] Step 21, calculate the moving speed of the face frame in the horizontal coordinate direction, then regard the moving trajectory of the face frame between two frames as uniform linear motion, and calculate the moving speed v of the face frame in the horizontal coordinate direction;

[0115] Step 22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v;

[0116] Step 23: Calculate the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

[0117] Preferably, in step 22, a speed limit area is predicted based on the lateral swing range of the real face, and the moving speed v is limited within the speed limit area.

[0118] Predicting the speed limit area according to the lateral swing range of the real face and limiting the moving speed v in the speed limit area specifically includes:

[0119] Assume that the initial position of the static face frame is x0, the width of the face frame is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w.

[0120] The area from the left extreme position of face swing to the right extreme position of face swing is set as the speed limit area; the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1;

[0121] When the face frame is in a stationary state, the initial position x0 is reset.

[0122] Preferably, the time difference between the predicted frame and the current frame is smaller than the time difference between the current frame and the past frame.

[0123] Since the device introduced in the second embodiment of the present invention is a device used to implement the method of the first embodiment of the present invention, those skilled in the art can understand the specific structure and deformation of the device based on the method introduced in the first embodiment of the present invention, so it is not described here in detail. All devices used in the method of the first embodiment of the present invention belong to the scope of protection of the present invention.

[0124] The present invention can effectively improve the problem of low fluency of face recognition interface under low frame rate by predicting the position of face frame and inserting frames. The linear interpolation method is used to predict the face frame. Under normal circumstances, the position of the face frame can be well predicted. The range that needs speed restriction is estimated according to the maximum swing amplitude of the real face, and the speed restriction is imposed on it. This method effectively improves the problem of face frame deviation from the real face when the face suddenly moves in the opposite direction, and improves the accuracy of predicting the position of the face frame. By filtering the large changes in the horizontal coordinate position of the face frame between the past frame and the current frame, the problem of face deviation from the real face when different faces are switched at a long distance is solved, and the robustness of the system is improved.

[0125] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0127] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0128] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0129] Although the specific implementation modes of the present invention are described above, those skilled in the art should understand that the specific implementation modes described are only illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A frame insertion method for face recognition tracking display, It is characterized in that include: Step 1: According to the result of face detection, the face frame position information and face detection time of the previous frame are stored and continuously updated; Step 2: Estimate the face frame position information of a predicted frame at least one specified time between the current frame and the future frame according to the face frame position information of the previous frame and the face detection time and the face frame position information of the current frame; Step 3: Display the face frame according to the face frame position information at the specified time.

2. The method according to claim 1, It is characterized in that The step 2 specifically includes: Step 21, calculate the moving speed of the face frame in the horizontal coordinate direction, then regard the moving trajectory of the face frame between two frames as uniform linear motion, and calculate the moving speed v of the face frame in the horizontal coordinate direction; Step 22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v; Step 23: Calculate the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

3. The method according to claim 2, Features: In the step 22, the speed limit area is predicted according to the lateral swing range of the real face, and the moving speed v is limited within the speed limit area.

4. The method according to claim 3, It is characterized in that Predicting the speed limit area according to the lateral swing range of the real face and limiting the moving speed v in the speed limit area specifically includes: Assume that the initial position of the static face frame is x0, the face frame width is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w. The left extreme position of face swing to the right extreme position of face swing is taken as the real face lateral swing range, and the area outside the real face lateral swing range is set as the speed limit area, and the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1; When the face frame is in a stationary state, the initial position x0 is reset.

5. The method according to claim 1, Features: The time difference between the predicted frame and the current frame is smaller than the time difference between the current frame and the past frame.

6. A frame insertion device for face recognition tracking display, It is characterized in that include: A storage and update module is used to store the face frame position information and face detection time of the previous frame and continuously update them according to the face detection result; A prediction frame module, used to estimate the face frame position information of a prediction frame of at least a specified time between the current frame and the future frame according to the face frame position information and face detection time of the previous frame and the face frame position information of the current frame; The display module is used to display the face frame according to the face frame position information at a specified time.

7. The device according to claim 6, It is characterized in that The prediction frame module is specifically used to perform the following steps: Step 21, calculate the moving speed of the face frame in the horizontal coordinate direction, then regard the moving trajectory of the face frame between two frames as uniform linear motion, and calculate the moving speed v of the face frame in the horizontal coordinate direction; Step 22, calculating the horizontal coordinate position of the face frame of the predicted frame according to the time difference between the predicted frame and the current frame and the moving speed v; Step 23: Calculate the vertical coordinate position of the face frame of the predicted frame according to the linear equation of the movement trajectory of the face frame of the past frame and the current frame and the horizontal coordinate position of the face frame of the predicted frame.

8. The device according to claim 7, Features: In the step 22, the speed limit area is predicted according to the lateral swing range of the real face, and the moving speed v is limited within the speed limit area.

9. The device according to claim 8, It is characterized in that Predicting the speed limit area according to the lateral swing range of the real face and limiting the moving speed v in the speed limit area specifically includes: Assume that the initial position of the static face frame is x0, the face frame width is w, and the width threshold is n. Then the left limit position of the face swing is x0-n×w, and the right limit position of the face swing is x0+n×w. The area from the left extreme position of face swing to the right extreme position of face swing is set as the speed limit area; the speed in the speed limit area is set to v×m, where m is the speed limit threshold and is less than 1; When the face frame is in a stationary state, the initial position x0 is reset.

10. The device according to claim 6, Features: The time difference between the predicted frame and the current frame is smaller than the time difference between the current frame and the past frame.