A Software Interaction Method and Device Based on Facial Recognition

Abstract

The present invention discloses a software interaction method and device based on face recognition. The method includes: the client acquires video and audio data containing a human face; extracts the action feature information of the key parts of the human face in the video data and the volume size information in the audio data; and converts the face action feature information and the volume size information into action instructions of virtual objects in the software through a body movement algorithm and a volume judgment algorithm. The software program of the present invention has high recognition accuracy for face actions, fast feedback speed, and integrates an SDK with a small volume and rich interfaces, which can be used by software engineers for subsequent optimization to continuously improve the user experience. On the other hand, users can conveniently and quickly exercise their eyes, facial muscles, and cervical vertebrae at any time and place through this software application, effectively preventing and treating the occupational diseases of computer users who work in the office.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent terminals, and in particular to a software interaction method and device based on face recognition. Background Art

[0002] In recent years, the public's attention to health has increased unprecedentedly, especially for "working people" struggling in the workplace. For the sake of happiness, many people are willing to risk their lives. Frequent health problems have also deepened the health awareness of workers.

[0003] The term occupational disease has a long history. The health problems of white-collar workers in the fast-paced workplace are hidden in the overtime work, the late nights and the takeaways they have ordered; "gradually receding hairline", "poop without fat", "mouse hand that can't hold the mouse", "deepening dark circles", "anxiety and pressure that always accompany me" and "belly that bulges out without knowing when" are all proof of hard work. In addition, according to a survey conducted by the Cervical Disease Prevention and Treatment Professional Committee of the China Medical and Health International Exchange Promotion Association, nearly 150 million people in my country have cervical problems; for white-collar workers born in the 1980s and 1985s, cervical and lumbar problems are particularly obvious, especially for legal and financial personnel.

[0004] For the vast majority of white-collar workers, most of the day is occupied by work, and it is unlikely that there is extra time to go to the gym or exercise outdoors. Therefore, a convenient exercise method and device are urgently needed to solve the above problems.

[0005] Most of the simple exercise devices and methods for one person on the market currently require activity space or additional equipment. The present invention only requires a mobile phone, which is a must-have for everyone in today's society. It does not require additional activity space and can be performed anytime and anywhere. Moreover, because it is an interactive game method, it is very interesting.

[0006] In summary, the present invention can meet the exercise needs of office workers and even people who have insufficient time and space conditions. It can effectively prevent and treat occupational diseases while bringing joy to users, and has broad application prospects. Summary of the invention

[0007] The present invention discloses a software interaction method and device based on face recognition, which comprises the following steps:

[0008] The extraction unit obtains video and audio data containing a human face;

[0009] The video is a complete action video containing all faces, and the audio data can be meaningful sentences or meaningless syllables.

[0010] The screening unit obtains the motion feature information of the key parts of the face in the video data by using the face recognition method;

[0011] The key facial parts are the specific facial features or areas for performing facial actions, such as eyes, nose, mouth, cheeks, etc. The above-mentioned face recognition method includes the following steps:

[0012] Obtain the average saturation value of each facial area image in the video data;

[0013] Perform an exponential function processing on the average saturation value of the facial area image to obtain the average saturation value of the key facial parts;

[0014] Compare the differences in the average saturation values of the key facial parts before and after the facial action and obtain the difference degree, which is the action feature information of the key facial parts.

[0015] The saturation of a color refers to the vividness of the color, also known as purity. In the color model, saturation is one of the three attributes of a color, and the other two attributes are hue and lightness; in this model, the value range of hue is from 0° to 360°, and the value ranges of saturation and lightness are from 0 to 100%. In color science, primary colors have the highest saturation. As the saturation decreases, the color becomes dull until it becomes achromatic, that is, the color loses its hue. As a carrier of information, color not only adheres to the design form but also acts as a subject to complete the process of information transmission. For image processing, the application and calculation of saturation are important parts. In practical applications, colors are divided into three basic categories of low-saturation contrast, medium-saturation contrast, and high-saturation contrast, as well as a derivative category of saturation combination contrast according to the high and low values of saturation.

[0016] The user's client should be equipped with an image processing program, which can read and transform the saturation of the target image, and can obtain the saturation value of each pixel in any area of the face image and calculate the average saturation value of the face area image according to the saturation values of each pixel.

[0017] Although the average saturation value of the obtained area image already contains the feature information of the key facial parts, the covered area range is too large. Therefore, this average saturation value is often inaccurate and needs to be further processed to obtain a more accurate result.

[0018] By performing exponential function processing on the average saturation value of the initially obtained face area image, the result of the average saturation value that only involves the key facial parts can be obtained.

[0019] The exponential function is represented by the mathematical model: y = k(3 x +7)

[0020] After obtaining the average saturation value of the key facial part image, divide the average saturation value of the image after the facial action by the average saturation value of the image before the facial action to obtain the difference degree, that is, the action feature information of the key facial parts.

[0021] The calculation unit converts the facial action feature information and audio data into action instructions for virtual objects in the software through the limb movement algorithm and the volume judgment algorithm.

[0022] Furthermore, in order to achieve the purpose of controlling virtual objects in the software through facial actions, a mapping relationship needs to be established between the action feature information of key facial parts and the movement of virtual objects in the software, which is realized through the limb movement algorithm at this time. The algorithm includes the following steps:

[0023] Divide the facial action feature information into head action feature information, eye action feature information, and mouth action feature information;

[0024] Construct a 3D facial model in the software environment according to a preset ratio;

[0025] Convert the head action feature information, eye action feature information, and mouth feature information into the coordinate offsets of the head key points, eye key points, and mouth key points in the virtual 3D facial features respectively;

[0026] Use the coordinate offsets of the head and eye key points as the animation displacement amount of the virtual object in the software, and send the coordinate offset of the mouth key point as the output result to the volume judgment algorithm.

[0027] The 3D facial model is built in the 3D software on the server side. The preset ratio is determined according to the actual needs during the development of the interactive software, rather than according to the ratio of the computer screen. If a larger displacement amount of the virtual object in the client video animation is required, the ratio is increased; if a relatively smaller displacement amount of the virtual object in the client video animation is desired, the ratio is decreased.

[0028] In order to achieve the purpose of controlling the size transformation of virtual objects through the volume of human voice, another algorithm, namely the volume judgment algorithm, needs to be embedded on the basis of face recognition.

[0029] Furthermore, the algorithm includes the following steps:

[0030] Preset the minimum time interval;

[0031] Perform an AND operation on the coordinate offset of the real-time mouth key point and the volume magnitude. If the result is 1, continue with the following steps; otherwise, terminate;

[0032] Use the volume change within the preset minimum time interval as the change amount of the size of the virtual object in the software environment.

[0033] When this algorithm is actually implemented in programming, an application programming interface access function needs to be added to the code to enable access to the API components of the local environment.

[0034] API, that is, Application Programming Interface, refers to some predefined interfaces (such as functions, HTTP interfaces), or the conventions for the connection between different components of a software system. It is used to provide a set of routines for applications and developers to access based on a certain software or hardware, without the need to access the source code or understand the details of the internal working mechanism.

[0035] After the volume judgment algorithm adds the API interface access function, it can conveniently call the audio device in the user's local device to execute the functions of volume input and processing.

[0036] On the other hand, another technical solution adopted by the present invention is: to provide a software interaction device based on face recognition, and the device includes:

[0037] Extraction unit: Extract video and audio data containing face actions through the client program;

[0038] Screening unit: Decompose the extracted video including face actions into images with different expressions and compare them to obtain the difference information of the color saturation of the key parts of the facial images under different expressions;

[0039] Calculation unit: Calculate the displacement and transformation amount of the virtual object in the interaction animation through the face action feature information and the volume change information;

[0040] Environment detection unit: When the brightness of the background environment where the client user is located is not sufficient to support the extraction unit to work, take remedial measures on the background environment so that the background environment can support the extraction unit to work.

[0041] Furthermore, the remedial measure of the environment detection unit is specifically: externally install a set of soft lights with a color rendering index greater than 95%, so that the soft lights cover the entire face area.

[0042] It can be seen from the above technical solutions that the present invention has at least the following advantages and positive effects compared with the prior art:

[0043] 1. Based on a large number of face data models for training and combined with AI algorithms, the present invention realizes accurate key point positioning, has high recognition accuracy, and thus achieves perfect fitting to the face and can recognize various postures and expressions.

[0044] 2. The present invention can provide acquisition schemes with different numbers of key points (5 points, 43 points, 81 points, 106 points, 300 points) according to the requirements of the application scenario, and can combine the prediction of the movement trend to quickly even super-real-time feedback the key point data.

[0045] 3. The present invention encapsulates an SDK with rich interfaces, and the SDK has the characteristics of a small installation package size and a small storage space occupation.

[0046] 4. Compared with traditional exercise methods, the present invention has the characteristics of low cost, low time and venue requirements. Users can take out their mobile phones at any time and anywhere, open the program to exercise.

[0047] 5. Compared with traditional exercise methods, the present invention is more interesting. The present invention allows users to interact with the video animation in the game through small games. Since the scenes in the game are constantly changing, users can appreciate the rich and colorful animation effects while putting themselves into the virtual characters in the game, and will not feel bored, thereby helping users develop a habit of long-term exercise.

[0048] 6. Different from traditional face recognition interactive programs, the present invention adopts an algorithm based on the recognition of changes in facial key point image features, which can effectively eliminate the impact of facial decorations or obstructions on accurate face recognition. Face sticker display scripts can also be added to subsequent software expansion packages to make the animation effects seen by users when interacting with the software richer.

[0049] 7. Different from the traditional face recognition interactive program, the expression detection function of the present invention is more powerful. In addition to the general blinking and pouting expressions, the present invention can recognize more complex and subtle expression changes, such as pursing the lips, biting the lips, puffing the cheeks, squinting or enlarging the eyes, etc., because it adopts the face recognition algorithm based on high-pixel image detail detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0051] Figure 1 It is a flowchart of a software interaction method based on face recognition provided by an embodiment of the present invention;

[0052] Figure 2 is a flow chart of a limb movement algorithm provided by another embodiment of the present invention;

[0053] Figure 3 is a flow chart of a volume determination algorithm provided by another embodiment of the present invention;

[0054] Figure 4 It is a structural diagram of a software interaction device based on face recognition provided by another embodiment of the present invention. DETAILED DESCRIPTION

[0055] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0056] Embodiment 1

[0057] Figure 1 It is a schematic flowchart of a software interaction method based on face recognition provided by an embodiment of the present invention. The method includes the following steps:

[0058] S1: The extraction unit acquires video and audio data containing a human face.

[0059] S2: The screening unit obtains the motion feature information of the key parts of the human face in the video data through the face recognition method.

[0060] S3: The calculation unit converts the face motion feature information and audio data into motion instructions for virtual objects in the software through the body motion algorithm and volume judgment algorithm.

[0061] Embodiment 2

[0062] Figure 2 It is a schematic flowchart of the body motion algorithm provided by another embodiment of the present invention. The method includes the following steps:

[0063] S30: Classify the face motion feature information into head motion feature information, eye motion feature information, and mouth motion feature information.

[0064] The office health care mini-program activates the shooting device of the client, monitors the head and face movements of the user, and captures the original video based on the face. Since the captured original video is a video of the whole face, and the user may only move a partial area of the face, it is necessary to perform regional detection on the face. The prerequisite for regional detection is to make a preliminary regional division of the face. This program first makes a preliminary division into six regions including each facial feature through the sample model in the face sample library. After locating the center point of the region, a detection radius is set with the center point of the region as the center; for example, when the obtained original face image is 100x100 pixels, an eye region is divided according to the sample model, and with the center of this region as the center, the detection radius is set to 5 pixels. By judging the discriminant method of face features, it is determined whether there are key feature points with high discriminability in this detection circle region, that is, the features of the key parts of the face. If no key feature points with high discriminability are found after judgment, the detection radius is gradually enlarged according to a certain progressive distance. For example, if the progressive distance is 1 pixel, the detection is carried out in the order of increasing detection radii of 6 pixels, 7 pixels, 8 pixels, etc. until the key feature points are detected, and then the detection is successful. In principle, the smaller the progressive distance, the better, because the smaller the distance, the higher the detection accuracy, and this distance can be set by the user himself. There are many methods for judging whether there are feature points on the face, such as the mosaic method, the eigenface method, the texture map method, the continuous symmetry method, the skin color region determination method, etc. After determining the smallest detection region, the video containing the face movement in this region is intercepted, that is, the movement video of the key parts of the face, and this video is divided into the head movement feature information video, the eye movement feature information video, and the mouth movement feature information video.

[0065] S31: Construct a 3D face model in the software environment according to a preset ratio;

[0066] The server side of the office health care mini-program constructs a virtual 3D face model according to the face image information captured by the client. This ratio is not constructed according to the display screen size of the user's client, but the ratio of the visual displacement of the virtual object moving in the office health care mini-program to the face in the image. If you want to increase the moving distance of the virtual object per unit time, then increase this ratio accordingly; if you want to decrease the moving distance of the virtual object per unit time, then decrease this ratio accordingly.

[0067] S32: Respectively use the head movement feature information, the eye movement feature information, and the mouth feature information as the coordinate offsets of the head key points, the eye key points, and the mouth key points in the virtual 3D facial features.

[0068] After the virtual human face 3D model is built, the server first receives the head motion feature information video, eye motion feature information video, and mouth motion feature information video transmitted from the client. Then, it extracts the head key point motion change video from the head motion feature information video, the eye key point motion change video from the eye motion feature information video, and the mouth key point motion change video from the mouth motion feature information video. Subsequently, the above three videos are applied to the built virtual human face model, which can be achieved by the server's own 3D animation generation function or by the server invoking external 3D animation software. The server records the changes in the coordinate values of these key points in the 3D human face model to obtain the specific coordinate displacement. The eye key points can be the center points of the upper eyelid, lower eyelid, eyelid, and eyeball, etc.; the head key points are a series of feature points that form the boundary line of the head contour; the mouth key points can be the center points of the upper lip and lower lip, etc.

[0069] S33: Use the coordinate offsets of the head key points and eye key points as the animation displacement of the virtual object in the software, and send the coordinate offset of the mouth key points to the volume judgment algorithm as the output result.

[0070] After the server obtains the coordinate offsets of the head key points and eye key points, it sends the offsets to the displacement variable of the virtual object used for user interaction in the client animation generation program. When the animation script calls the virtual object and outputs it to the display screen of the user, the position of the virtual object will change in real time with the displacement variable.

[0071] When the user selects the eye care option in the health care program, after the program starts, as the user blinks, the virtual character in the animation will make a bouncing motion at the same time as the user blinks, so as to avoid the roadblocks on the ground in the animation. Since the roadblocks are continuous, the user needs to keep blinking, otherwise, if hitting the roadblock, the animation will terminate and the scenery on both sides of the roadblock in the subsequent animation cannot be viewed. When the user feels eye fatigue due to long-term computer gazing, they can pick up their mobile phone to start the health care program during the break time of work to relieve eye fatigue.

[0072] When the user selects the head health care option in the health care program, after the program starts, directional arrows will fly out horizontally from left to right on the screen. When the directional arrows reach the virtual box on the right side of the screen, the user needs to make corresponding neck-turning movements upward, downward, leftward, and rightward according to the directions indicated by the directional arrows. If the correct neck-turning movement in the correct direction is made in time, one point is scored; if the correct movement in the correct direction is not made in time, no score is given. After the animation ends, the user will get the total score for this round. The user can improve the score by using this program multiple times, refresh their score record, and conveniently achieve the purpose of cervical spine health care while obtaining a sense of achievement.

[0073] Embodiment III

[0074] Figure 3 It is a flow schematic diagram of the volume judgment algorithm provided for another embodiment of the present invention. The method includes the following steps:

[0075] S34: Preset the minimum time interval;

[0076] The balloon-blowing mini-game first prompts the user to set the minimum detection time interval for blowing detection. The game program can divide the difficulty of the game according to the detected time interval range, which can be divided into primary, intermediate, advanced, etc. Subsequently, the game program obtains the volume data within the preset time interval by activating the audio input device of the user terminal.

[0077] S35: Perform an AND operation on the coordinate offset of the real-time mouth key point and the volume size. If the result is 1, continue with the following steps, otherwise terminate;

[0078] The game program obtains the real-time coordinate offset of the mouth key point through the limb movement algorithm on the server side. If the offset is greater than zero, it proves that the user's mouth is moving at this time. Then, it is judged whether the volume is greater than zero at this time. If the volume is greater than zero and the coordinate offset of the mouth key point is also greater than zero, it proves that the user is blowing and making a sound at this time, rather than background noise; the above judgment method can use an AND operation in logical operations. Only when the coordinate offset of the mouth key point is greater than zero, the value of the coordinate offset is 1 in logical operations, and the volume is also greater than 0, and the value of the volume size is also 1 in logical operations, the result of the logical AND operation of the coordinate offset of the mouth key point and the volume size is 1, otherwise it cannot be judged whether it is a person making a sound, thus affecting the progress of the game.

[0079] S36: Use the volume change within the preset minimum time interval as the change amount of the size of the virtual object in the software environment.

[0080] In the balloon-blowing mini-game, the user can affect the degree of balloon inflation in the game animation by continuously increasing the blowing volume and the vocal volume. The longer the continuous blowing time and the louder the volume, the greater the degree of balloon inflation. To let the user experience the fun of playing the game, different phased goals can be added to the game. Different levels can be set up, and a maximum inflation degree is set for the balloon in each level. When the user's vocal volume exceeds this inflation degree, the virtual balloon will burst, and at the same time, an animation after bursting is added, along with a reward prompt to encourage the user to challenge a greater blowing volume and keep passing levels. When the user is playing the balloon-blowing mini-game, they exercise their lung capacity and also achieve the purpose of physical and mental pleasure, which can be described as killing two birds with one stone.

[0081] Embodiment 4

[0082] Figure 4 The figure is a schematic structural diagram of a software interaction device based on face recognition provided by another embodiment of the present invention. The device includes an extraction unit 100, a screening unit 200, a calculation unit 300, and an environment detection unit 400. Among them, the extraction unit 100 obtains video and audio data containing human face movements, which can be through the camera of a mobile phone, the camera of a computer, or other video shooting and monitoring devices.

[0083] The screening unit 200 is used to receive the face video obtained by the extraction unit 100, screen out the action feature information of the key parts of the face from it, and send it to the calculation unit 300; the calculation unit 300 converts the action feature information of the key parts of the face into the movement and transformation amount of the virtual object on the user client side; the above process is for the case where the background light source is sufficient. When the number of background light sources is not enough for the extraction unit to extract the face movements or expression changes from the captured video of the face to be detected, the environment detection unit 400 should promptly notify the user to add an external light source, and the color rendering index of this light source should be above 95% so as not to lose any necessary details of the face features. If the added external light source meets the requirements, the user will be prompted "Environmental light repair successful" through the client, otherwise, it will prompt "Color rendering index not up to standard" or "Environmental illuminance not up to standard".

[0084] The above are only the embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be included in the patent protection scope of the present invention by the same token.

Claims

1. A software interaction method based on face recognition, characterized in that, It includes the following steps: S1: The extraction unit acquires video and audio data containing human faces; S2: The screening unit obtains the action feature information of the key parts of the human face in the video data through the face recognition method; among them, obtaining the action feature information of the key parts of the human face in the video data specifically includes: Dividing the original human face image in the video data into six regions including facial features through the sample model in the human face sample library; Locating the center point of the region, taking the center point of the region as the center of the circle, and setting the detection radius; Judging whether the features of the key parts of the human face are contained in the current detected circular region through the discrimination method of human face features; If the features of the key parts of the human face cannot be found, the detection radius is enlarged according to the preset progressive distance until the features of the key parts of the human face are detected, the detection is successful, and the minimum detection region is obtained; After determining the minimum detection region, intercept the video containing the human face action in the current detection region to obtain the action feature information of the key parts of the human face; S3: The calculation unit converts the action feature information of the key parts of the human face and the audio data into action instructions of virtual objects in the software through the body movement algorithm and the volume judgment algorithm; Among them, the body movement algorithm includes the following steps: S30: Divide the action feature information of the key parts of the human face into head action feature information, eye action feature information, and mouth action feature information; S31: Build a 3D human face model in the software environment according to the preset ratio; among them, the ratio is preset according to the ratio of the visual displacement amount of the virtual object movement to the ratio of the human face in the image; S32: Convert the head action feature information, eye action feature information, and mouth action feature information into the coordinate offsets of the head key points, eye key points, and mouth key points in the virtual 3D facial features of the human face respectively; S33: Use the coordinate offsets of the head key points and eye key points as the animation displacement amounts of the virtual objects in the software, and send the coordinate offsets of the mouth key points as the output results to the volume judgment algorithm; Among them, the volume judgment algorithm includes the following steps: S34: Preset the minimum time interval; S35: Perform an AND operation on the real-time coordinate offset of the mouth key point and the volume size. If the result is 1, continue with the following steps, otherwise terminate; S36: Take the volume change within the preset minimum time interval as the change amount of the size of the virtual object in the software environment.

2. The method according to claim 1, wherein the face recognition method in step S2 is characterized in that It includes the following steps: S20: Obtain the average saturation value of each facial region image of the human face in the video data; S21: Perform exponential function processing on the average saturation value of the facial region image to obtain the average saturation value of the key parts of the human face; S22: Compare the difference in the average saturation value of the key parts of the human face before and after the human face action and obtain the difference degree, and this difference degree is the action feature information of the key parts of the human face.

3. The method according to claim 2, wherein the exponential function in step S21 is characterized in that It is expressed by the following formula: y = k(3 x + 7); Among them, k is the required coefficient, and the value of k is a number between (0, 0.8); x is the average saturation value in each facial region image of the human face collected, and y is the average saturation value of the image of the key parts of the human face.

4. The method according to claim 2, wherein the difference degree described in step S22 is characterized in that The difference degree is the ratio of the average saturation value of the key part image after the facial action to the average saturation value of the key part image before the facial action.

5. A software interaction device based on face recognition, characterized in that, Including: Extraction unit: Extract video and audio data containing facial actions through the client program; Screening unit: Decompose the extracted video including facial actions into images with different expressions and compare them to obtain the difference information of the color saturation of the key parts of the facial images under different expressions; Among them, before obtaining the difference information of the color saturation of the key parts of the facial images under different expressions, it also includes: Divide the original facial image in the video data into six regions including facial features through the sample model in the facial sample library; Locate the center point of the region, and set the detection radius with the center point of the region as the center; Judge whether the features of the key parts of the face are contained in the current detected circular region through the discrimination method of facial features; If the features of the key parts of the face cannot be found, expand the detection radius according to the preset progressive distance until the features of the key parts of the face are detected, the detection is successful, and the minimum detection region is obtained; After determining the minimum detection region, intercept the video containing facial actions in the current detected region to obtain the action feature information of the key parts of the face; Calculation unit: Calculate the displacement and transformation amount of the virtual object in the interactive animation through the action feature information of the key parts of the face and the volume change information; Among them, calculating the displacement and transformation amount of the virtual object in the interactive animation is realized through the limb movement algorithm and the volume judgment algorithm; Among them, the limb movement algorithm includes the following steps: Divide the action feature information of the key parts of the face into head action feature information, eye action feature information, and mouth action feature information; Construct a 3D facial model in the software environment according to the preset ratio; Among them, the ratio is preset according to the ratio of the visual displacement amount of the virtual object movement to the ratio of the face in the image; Respectively convert the head action feature information, eye action feature information, and mouth action feature information into the coordinate offsets of the head key points, eye key points, and mouth key points in the virtual 3D facial features; Use the coordinate offsets of the head key points and eye key points as the animation displacement amount of the virtual object in the software, and send the coordinate offset of the mouth key point as the output result to the volume judgment algorithm; Among them, the volume judgment algorithm includes the following steps: Preset the minimum time interval; Perform an AND operation on the coordinate offset of the real-time mouth key point and the volume size. If the result is 1, continue with the following steps, otherwise terminate; Use the volume change within the preset minimum time interval as the change amount of the size of the virtual object in the software environment; Environmental detection unit: When the brightness of the background environment where the client user is located is not sufficient to support the extraction unit to work, take remedial measures on the background environment so that the background environment can support the extraction unit to work.

6. The remedial measure of the environmental detection unit of the device according to claim 5, characterized in that Externalize a set of soft lights with a color rendering index greater than 95% so that the soft lights cover the entire facial area.

7. A client, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it realizes the steps of the method according to any one of claims 1 to 4.

8. A computer-readable storage medium storing a computer program and applied to a client, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

Images