A computer performance adjustment method and system based on user facial expression recognition

Abstract

The application provides a computer performance adjustment method and system based on user expression recognition, relates to the technical field of computer performance adjustment, and solves the problem of a large number of external devices required in the current user expression recognition and performance adjustment method, interference with normal use of the computer by the user, and poor user experience. First, an expression picture of the user using the computer is acquired, then the expression picture of the user using the computer is recognized to obtain a user emotional expression, then the change of the user emotional expression is judged according to the obtained user emotional expression, then the user experience is judged by using the change of the user emotional expression, and finally the computer performance is dynamically adjusted according to the user experience until the computer performance reaches the best, without the need of connecting a large number of external devices, so that the user can normally use the computer and the user experience is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer performance adjustment, and particularly relates to a computer performance adjustment method and system based on user expression recognition. BACKGROUND

[0002] With the vigorous development of computer technology, human society has moved from a highly industrialized era to a computer era, and computers have entered various fields of people's lives and are closely related to people's daily work, study and life.

[0003] At present, computer technology has developed rapidly in terms of operation speed, performance expansion, efficient power consumption and application development, and the standards of computer users' operation requirements for computers are gradually increasing. Therefore, computers need to dynamically meet the performance requirements of computer users by adjusting the clock frequency of the CPU, the large core and small core of the CPU, and the clock frequency of the GPU. The performance adjustment method of the traditional computer mainly includes a simple or "big move" performance adjustment method. The simple performance adjustment method refers to adding an external mechanical device, and the user inputs user experience information into the computer through the external mechanical device. The computer dynamically adjusts the computer performance according to the user experience, but this method greatly interferes with the normal use of the computer by the user. The "big move" performance adjustment method refers to using a large number of external sensing devices to obtain user biometric information, and then predicting user experience through the biometric information, and dynamically adjusting the computer performance according to the user experience. However, this method requires too many external sensing devices, and the external sensing devices also consume power, which is high in cost. Both of the above-mentioned traditional computer performance adjustment methods are associated with user experience and computer performance, and ensure user experience in the process of adjusting computer performance. A performance adjustment method is also disclosed in the prior art, which updates the pre-designed computer performance parameters according to the adjustment parameters of the performance parameter adjustment operation, so as to run the game program with the updated performance parameters. However, the computer performance parameters are adjusted by relying on the user input performance parameter adjustment operation to meet the user experience, and the situation that the performance parameter adjustment process will cause poor user experience is ignored, which affects the user experience. SUMMARY

[0004] To solve the problem of the current computer performance adjustment method, which requires a large number of external devices, interferes with the normal use of the computer by the user, and has poor user experience, the present application provides a computer performance adjustment method and system based on user expression recognition, which does not require a large number of external devices, ensures the normal use of the computer by the user, and guarantees the user experience.

[0005] In order to achieve the above technical effects, the technical scheme of the present application is as follows:

[0006] A computer performance adjustment method based on user expression recognition, comprising the following steps:

[0007] S1. Obtain an expression picture of a user when using a computer;

[0008] S2. Recognize the expression picture to obtain a user emotional expression;

[0009] S3. Determine a change in the user emotional expression according to the obtained user emotional expression;

[0010] S4. Determine a user experience by using the change in the user emotional expression;

[0011] S5. Dynamically adjust a computer performance according to the user experience until the computer performance reaches an optimal state.

[0012] In the technical solution, the obtained expression picture is first recognized, then the change in the user emotional expression is used to determine the user experience, the change in the user emotional expression is associated with the user experience, the problem that the instantaneous user emotional expression cannot reflect the user experience is solved, the computer performance is dynamically adjusted according to the user experience, the user experience is associated with the computer performance, the situation that the user experience is poor due to the process of ignoring the computer performance is avoided, the purpose of balancing and coordinating the user experience and the computer performance adjustment is achieved, and finally the computer performance reaches the optimal state and stops adjusting. The technical solution does not need to access a large number of external devices, can ensure that the user normally uses the computer, and guarantees the user experience.

[0013] Preferably, in step S1, the expression picture is obtained by using a camera of the computer.

[0014] Preferably, in step S2, the expression picture is recognized by using a facial expression recognition technology, and the specific recognition steps are as follows: first, input the expression picture into a constructed convolutional neural network, then determine a quantization value corresponding to the expression picture, and finally output a user emotional expression corresponding to the quantization value by the convolutional neural network.

[0015] Preferably, the user emotional expression includes eight user emotional expressions of happy, contempt, sad, disgusted, surprised, angry, frightened and indifferent.

[0016] Preferably, in step S3, the specific steps of determining the change in the user emotional expression are as follows:

[0017] S31. Each user emotional expression is a random variable, and multiple user emotional expressions correspond to multiple random variables.

[0018] S32. Determine whether two random variables in adjacent recognition time are equal, if yes, the user emotional expression is stable, otherwise, the user emotional expression changes.

[0019] Preferably, in step S4, the specific step of judging the user experience is:

[0020] S41. Representing a plurality of random variables representing user emotional expressions as an overall quantity Δ representing the change of user emotional expressions face , the specific expression of the overall quantity Δ face is:

[0021] Δ face = δ h + δ c + δ sa + δ d + δ su + δ a + δ f + δ n

[0022] Wherein, δ h represents the result of the rank-sum test of the happy user emotional expression, δ c represents the result of the rank-sum test of the contemptuous user emotional expression, δ sa represents the result of the rank-sum test of the sad user emotional expression, δ d represents the result of the rank-sum test of the disgusted user emotional expression, δ su represents the result of the rank-sum test of the surprised user emotional expression, δ a represents the result of the rank-sum test of the angry user emotional expression, δ f represents the result of the rank-sum test of the fearful user emotional expression, and δ n represents the result of the rank-sum test of the indifferent user emotional expression.

[0023] S42. Setting an emotional expression change threshold, judging whether the overall quantity Δ face is greater than the emotional expression change threshold, if yes, the user experience changes; otherwise, the user experience remains stable.

[0024] Preferably, in step S41, the user emotional expression includes happy, contempt, sad, disgust, surprise, anger, fear and indifference, if the happy user emotional expression changes, the value of the result of the rank sum test corresponding to the happy user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the happy user emotional expression is 0; if the contempt user emotional expression changes, the value of the result of the rank sum test corresponding to the contempt user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the contempt user emotional expression is 0; if the sad user emotional expression changes, the value of the result of the rank sum test corresponding to the sad user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the sad user emotional expression is 0; if the disgust user emotional expression changes, the value of the result of the rank sum test corresponding to the disgust user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the disgust user emotional expression is 0; if the surprise user emotional expression changes, the value of the result of the rank sum test corresponding to the surprise user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the surprise user emotional expression is 0; if the anger user emotional expression changes, the value of the result of the rank sum test corresponding to the anger user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the anger user emotional expression is 0; if the fear user emotional expression changes, the value of the result of the rank sum test corresponding to the fear user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the fear user emotional expression is 0; if the indifference user emotional expression changes, the value of the result of the rank sum test corresponding to the indifference user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the indifference user emotional expression is 0.

[0025] Preferably, in the process of identifying the user emotional expression, noise interfering with the user emotional expression is set.

[0026] Preferably, in step S5, the specific operation steps of dynamically adjusting the computer performance are:

[0027] S51. Set the initial value M of the computer performance;

[0028] S52. Reduce the computer performance by P starting from the initial value M of the computer performance in step S51, judge whether the user experience changes, if yes, take the computer performance M corresponding to the change of the user experience as the best performance; otherwise, continue to reduce the computer performance by P, execute step S53;

[0029] S53. Determine whether the user experience changes, if yes, set the computer performance corresponding to the user experience change as a, take P as the computer performance change, increase the computer performance a corresponding to the user experience change by P to obtain the computer performance P+a corresponding to the user experience change before the change, and take the computer performance P+a as the best performance; otherwise, perform step S54;

[0030] S54. Continue to reduce the computer performance by P, and return to S53.

[0031] The application also proposes a computer performance adjustment system for user expression recognition, the system comprising:

[0032] An acquisition module for acquiring expression pictures of a user using a computer;

[0033] An identification module for identifying the expression pictures to obtain user emotional expressions;

[0034] A first determination module for determining the change of the user emotional expressions according to the obtained user emotional expressions;

[0035] A second determination module for determining the user experience by using the change of the user emotional expressions;

[0036] An adjustment module for dynamically adjusting the computer performance according to the user experience until the computer performance reaches the best.

[0037] Preferably, in the identification module, the user emotional expressions include happy, contempt, sad, disgust, surprise, anger, fear and indifference.

[0038] Compared with the prior art, the technical scheme of the application has the beneficial effects that:

[0039] The application proposes a computer performance adjustment method and system based on user expression recognition, in which the acquired expression pictures are first identified, then the change of the user emotional expressions is used to determine the user experience, the change of the user emotional expressions and the user experience are associated, the problem that the instantaneous user emotional expressions cannot reflect the user experience is solved, the computer performance is dynamically adjusted according to the user experience, the user experience and the computer performance are associated, the situation that the user experience is poor due to the process of ignoring the computer performance is avoided, the purpose of balancing and coordinating the user experience and the computer performance adjustment is achieved, and finally the computer performance reaches the best and stops adjusting, the technical scheme does not need to access a large number of external devices, can ensure the normal use of the computer by the user, and guarantees the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1Fig. 1 shows a flow chart of a computer performance adjustment method based on user facial expression recognition according to an embodiment of the present application;

[0041] Figure 2 Fig. 2 shows a change curve of a happy facial expression of a user in a high performance mode and a low performance mode;

[0042] Figure 3 Fig. 3 shows a change curve of a contempt facial expression of a user in a high performance mode and a low performance mode;

[0043] Figure 4 Fig. 4 shows a change curve of a sad facial expression of a user in a high performance mode and a low performance mode;

[0044] Figure 5 Fig. 5 shows a change curve of a disgust facial expression of a user in a high performance mode and a low performance mode;

[0045] Figure 6 Fig. 6 shows a change curve of a surprise facial expression of a user in a high performance mode and a low performance mode;

[0046] Figure 7 Fig. 7 shows a change curve of an angry facial expression of a user in a high performance mode and a low performance mode;

[0047] Figure 8 Fig. 8 shows a change curve of a fear facial expression of a user in a high performance mode and a low performance mode;

[0048] Figure 9 Fig. 9 shows a change curve of a neutral facial expression of a user in a high performance mode and a low performance mode;

[0049] Figure 10 Fig. 10 shows a structure diagram of a computer performance adjustment system based on user facial expression recognition according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] The accompanying drawings are only used for illustrative description and cannot be understood as limitation to the present patent;

[0051] In order to better illustrate the embodiments, some parts of the drawings are omitted, enlarged or reduced, and the description of "up", "down", etc. is not limitation to the present patent;

[0052] For those skilled in the art, it is understandable that some known content in the drawings can be omitted;

[0053] The position relationship described in the drawings is only used for illustrative description and cannot be understood as limitation to the present patent;

[0054] The technical solutions of the present application will be further described in combination with the drawings and embodiments.

[0055] Embodiment 1

[0056] As Figure 1 shown, the embodiment proposes a computer performance adjustment method based on user expression recognition, comprising the following steps:

[0057] S1. Obtain an expression picture of a user using a computer;

[0058] In step S1, the expression picture is obtained by using the camera of the computer. When actually using the computer, the camera of the computer is at the best angle for observing the user's expression, and the facial expression of the computer user is photographed or filmed. If it is photographed, the expression picture of the user using the computer can be directly obtained, and if it is filmed, the expression picture of the user using the computer needs to be obtained from the filmed video of the facial expression of the computer user. When the face of the computer user is not in the shooting range of the camera of the computer, the computer suspends photographing or filming the facial expression of the computer user until the face of the computer user appears in the shooting range of the camera of the computer.

[0059] S2. Identify the expression picture to obtain a user emotional expression;

[0060] In step S2, the expression picture is identified by using facial expression recognition technology. Facial expression recognition technology refers to separating a specific expression state from a given static image or dynamic video sequence to determine the psychological emotion of the identified object, realize the understanding and recognition of the computer to the facial expression, fundamentally change the relationship between people and computers, and achieve better human-computer interaction. The specific identification steps are as follows: first, input the expression picture into the constructed convolutional neural network, then determine the quantization value corresponding to the expression picture, and finally output the user emotional expression corresponding to the quantization value from the convolutional neural network; each expression picture corresponds to an eight-tuple, and each element in the eight-tuple corresponds to a basic user emotional expression, and the quantization value range of each user emotional expression is generally from 0 to 1; the user emotional expression includes happy, contempt, sad, disgust, surprise, anger, fear and indifference, and the above eight user emotional expressions are the results of modern psychological research.

[0061] S3. Determine the change of the user emotional expression according to the obtained user emotional expression;

[0062] In step S3, the specific steps for determining the change of the user emotional expression are as follows:

[0063] S31. Let each user emotional expression be a random variable, then a plurality of user emotional expressions correspond to a plurality of random variables;

[0064] S32. Determine whether two random variables in adjacent recognition time are equal, if yes, the user emotional expression is stable; otherwise, the user emotional expression changes.

[0065] S4. Determine the user experience by the change of the user emotional expression.

[0066] In step S4, people intuitively think that happy emotional expression means better user experience, and unhappy emotional expression means worse user experience, but the actual situation does not conform to people's intuitive judgment, and in most cases, happy emotional expression means worse user experience, and contempt, sad, disgusted, surprised, angry, frightened and indifferent emotional expression means better user experience. The above shows that the instantaneous user emotional expression is difficult to reflect the user experience. In addition, existing neurology research has proved that external stimulation can cause changes in user expression, and this external stimulation can be a picture, an animal, or user experience. Therefore, determining the user experience by the change of the user emotional expression relates the change of the user emotional expression to the user experience.

[0067] S5. Dynamically adjust the computer performance according to the user experience until the computer performance reaches the best.

[0068] Embodiment 2

[0069] In step S4, the specific steps for determining the user experience are:

[0070] S41. Express a plurality of random variables representing the user emotional expression as an overall quantity Δ representing the change of the user emotional expression face , the specific expression of the overall quantity Δ face is:

[0071] Δ face = δ h + δ c + δ sa + δ d + δ su + δ a + δ f + δ n

[0072] Wherein, δ h represents the result of the rank sum test of happy user emotional expression, δ c represents the result of the rank sum test of contempt user emotional expression, δ sa represents the result of the rank sum test of sad user emotional expression, δ d represents the result of the rank sum test of disgusted user emotional expression, δ su represents the result of the rank sum test of surprised user emotional expression, and δ athe result of the rank-sum test representing the user emotional expression of contempt, δ f the result of the rank-sum test representing the user emotional expression of fear, δ n the result of the rank-sum test representing the user emotional expression of indifference, δ

[0073] In step S41, the user emotional expressions include happiness, contempt, sadness, disgust, surprise, anger, fear, and indifference. If the user emotional expression of happiness changes, the value of the result of the rank-sum test corresponding to the user emotional expression of happiness is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of happiness is 0. If the user emotional expression of contempt changes, the value of the result of the rank-sum test corresponding to the user emotional expression of contempt is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of contempt is 0. If the user emotional expression of sadness changes, the value of the result of the rank-sum test corresponding to the user emotional expression of sadness is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of sadness is 0. If the user emotional expression of disgust changes, the value of the result of the rank-sum test corresponding to the user emotional expression of disgust is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of disgust is 0. If the user emotional expression of surprise changes, the value of the result of the rank-sum test corresponding to the user emotional expression of surprise is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of surprise is 0. If the user emotional expression of anger changes, the value of the result of the rank-sum test corresponding to the user emotional expression of anger is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of anger is 0. If the user emotional expression of fear changes, the value of the result of the rank-sum test corresponding to the user emotional expression of fear is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of fear is 0. If the user emotional expression of indifference changes, the value of the result of the rank-sum test corresponding to the user emotional expression of indifference is 1, otherwise, the value of the result of the rank-sum test corresponding to the user emotional expression of indifference is 0.

[0074] S42. Set an emotional expression change threshold, judge whether the overall quantity Δ face is greater than the emotional expression change threshold. If yes, the user experience changes; otherwise, the user experience remains stable. The specific formula is as follows:

[0075] UX_is_changed = Δ face > threshold

[0076] wherein UX_is_changed represents the prediction result of the change of the user emotional expression.

[0077] In this embodiment, a user is selected to make observation of the change of the user's emotional expression in the high performance and low performance modes of the computer respectively, the value of the change threshold threshold is set to 2, in the high performance mode, the performance of the computer is set to the highest, the computer runs very smoothly, which brings the best user experience to the user; in the low performance mode, the performance of the computer is set to the lowest, at this time the computer will have obvious lag, which brings the user a bad user experience, see Figures 2 to 9 For this user, the facial expression will show "happy" in the case of good user experience, it is worth noting that the user's emotional expression shows significant difference in the best user experience and the bad user experience; this embodiment further selects 20 users to make the same test, the test result shows that different users will show different facial expressions for good user experience, for example, for good user experience, some users will mainly show "happy", and some users will mainly show "indifference"; for bad user experience, different users also have different expressions, for bad user experience, some users will show "happy", and some users will show "indifference", the above results can verify that the user's instantaneous expression is difficult to judge the user experience.

[0078] Embodiment 3

[0079] See Figure 1 In step S5, when the user actually uses the computer, there will be many unavoidable noise to interfere with the user's emotional expression, these noises will interfere with the judgment of the change of the user's emotional expression, and further affect the judgment of the user experience, therefore, this embodiment is provided in the process of identifying the user's emotional expression, there are noises interfering with the user's emotional expression, these noises include the influence of computer application itself rather than computer hardware performance on expression, for example, when the user is immersed in the game, sometimes happy, sometimes frustrated, which will interfere with the user's expression, therefore, this embodiment designs a delicate algorithm to dynamically adjust the computer performance, the specific operation steps of dynamically adjusting the computer performance are:

[0080] S51. Set the initial value M of the computer performance;

[0081] In step S51, the higher the performance of the computer, the better the user experience, the computer performance is set at the point of the best user experience, the highest computer performance brings the best user experience, by starting the CPU large core and small core at the same time, setting the frequency to the highest, setting the frequency of the GPU to the highest, setting the computer performance to the highest, and setting the initial value M of the current computer performance;

[0082] S52. With P as the computer performance change amount, the computer performance is reduced by P from the initial value M of S51, it is judged whether the user experience is changed, if yes, the computer performance corresponding to the user experience before the change is taken as the best performance; otherwise, the computer performance is continuously reduced by P with P as the computer performance change amount, and S53 is executed;

[0083] In S52, the computer performance change amount P is a very small change amount, and P can be set with different values in actual performance adjustment.

[0084] S53. It is judged whether the user experience is changed, if yes, the computer performance corresponding to the user experience when the change occurs is taken as a, the computer performance corresponding to the user experience before the change is obtained by increasing a by P with P as the computer performance change amount, and the computer performance P+a is taken as the best performance; otherwise, S54 is executed;

[0085] S54. The computer performance is continuously reduced with P as the computer performance change amount, and S53 is returned.

[0086] The algorithm steps of dynamically adjusting the computer performance can tolerate and overcome the noises, and the algorithm can find the "best" performance of the computer without being affected by the noises, which can complete the computing task with minimum energy consumption and meet the "best" user experience. The algorithm of dynamically adjusting the computer performance of the embodiment is long-term running in the background of the computer, so as to adjust the computer performance in real time according to the user experience, achieve the purpose of balancing and coordinating the user experience and the computer performance adjustment, and realize the goal of efficient use of electric energy. The energy saving effect of the technical scheme of the embodiment is verified by experiments on three Samsung S20, S5 and Note4 smart phones. Three application scenarios are selected, which are typing, playing games and browsing web pages. Twenty users are recruited to use the three smart phones to complete the three application scenarios. The ondemand strategy of the Linux operating system is used as a benchmark for comparison. The experimental results show that the algorithm of the embodiment saves 17.3% of the minimum power and 52.9% of the maximum power compared with the ondemand algorithm. Five users are selected to score the user experience of the embodiment and two existing technologies. The five users use the scheme of the embodiment, the simplest scheme and the maximum interference scheme on a Macbook notebook computer for experiments. The simplest scheme is to add external mechanical devices, and the user can directly tell the computer whether the user is satisfied or not satisfied with the computer performance through the external mechanical devices. If not satisfied, the computer improves the performance, otherwise, the performance is appropriately reduced to save the electric energy. The maximum interference scheme is to use more external sensing devices to obtain the biological feature information of the user, and then predict the satisfaction of the user to the computer performance through the information. Five users play games under the three schemes, and then score the user experience. The score is the highest 5 points and the lowest 1 point. Among them, 5 points represent the most satisfied, 4 points represent general satisfaction, 3 points represent acceptable, 2 points represent not very satisfied, and 1 point represent very dissatisfied. Referring to Table 1, the average score of the scheme of the embodiment is 4.5 points, the average score of the simplest scheme is 1.2 points, and the average score of the maximum interference scheme is 2.8 points. From the scores, it can be known that the user experience of the scheme of the embodiment is the best.

[0087] Table 1: User experience score table of three schemes

[0088]

[0089] Embodiment 4

[0090] Referring to Figure 10 The embodiment proposes a computer performance adjustment system based on user expression recognition, which comprises:

[0091] The acquisition module is configured to acquire an expression picture of a user using a computer.

[0092] The user emotional expressions in the recognition module include happy, contempt, sad, disgust, surprise, anger, fear and indifference.

[0093] The recognition module is used for recognizing the expression picture to obtain user emotional expressions.

[0094] The first judging module is used for judging the change of the user emotional expressions according to the obtained user emotional expressions.

[0095] The second judging module is used for judging the user experience by using the change of the user emotional expressions.

[0096] The adjusting module is used for dynamically adjusting the computer performance according to the user experience until the computer performance reaches the best.

[0097] In the embodiment, the obtained expression picture is first recognized, then the change of the user emotional expressions is used to judge the user experience, the change of the user emotional expressions and the user experience are associated, the problem that the instantaneous user emotional expressions cannot reflect the user experience is solved, then the computer performance is dynamically adjusted according to the user experience, the computer performance and the user experience are associated, the situation that the user experience is poor caused by ignoring the computer performance is avoided, the purpose of balancing and coordinating the user experience and the computer performance adjustment is achieved, finally the computer performance reaches the best and stops adjusting, the technical scheme does not need to access a large number of external devices, the user can normally use the computer, and the user experience is ensured.

[0098] Obviously, the above embodiments of the present application are only examples for clearly illustrating the present application, and are not intended to limit the implementation manners of the present application. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. All the implementation manners do not need to be exhausted here. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the claims of the present application.

Claims

1. A computer performance adjustment method based on user expression recognition, characterized by, The method comprises the following steps: S1. obtaining an expression picture of a user when using a computer; S2. identifying the expression picture to obtain a user emotional expression; S3. judging a change in the user emotional expression according to the obtained user emotional expression; S4. judging a user experience by using the change in the user emotional expression; S5. dynamically adjusting computer performance according to the user experience until the computer performance reaches an optimum; In step S5, the specific operation steps of dynamically adjusting the computer performance are as follows: S51. setting an initial value M of the computer performance, including setting the computer performance at a point of optimum user experience, the highest computer performance bringing the best user experience, setting the computer performance at the highest by simultaneously starting CPU large cores and small cores, setting the frequency at the highest, setting the frequency of the GPU at the highest, and setting the initial value M of the current computer performance; S52. reducing the computer performance by P from the initial value M of step S51, judging whether the user experience changes, if yes, taking the computer performance corresponding to the user experience before the change as the optimum performance, otherwise, continuing to reduce the computer performance by P, and executing step S53; S53. judging whether the user experience changes, if yes, taking the computer performance corresponding to the user experience when the change occurs as a, increasing the computer performance corresponding to the user experience when the change occurs by P to obtain the computer performance corresponding to the user experience before the change, and taking the computer performance P+a as the optimum performance; otherwise, executing step S54; S54. continuing to reduce the computer performance by P, and returning to S53.

2. The computer performance adjustment method based on user facial expression recognition according to claim 1, characterized in that, In step S1, the expression picture is obtained by using a camera of the computer. 3.The computer performance adjustment method based on user facial expression recognition of claim 1, wherein, In step S2, the expression picture is identified by using a facial expression recognition technology, and the specific identification steps are as follows: first, inputting the expression picture into a constructed convolutional neural network, then determining a quantization value corresponding to the expression picture, and finally outputting a user emotional expression corresponding to the quantization value from the convolutional neural network.

4. The computer performance adjustment method based on user facial expression recognition according to claim 3, characterized in that, The user emotional expression includes eight user emotional expressions of happy, contempt, sad, disgusted, surprised, angry, frightened, and indifferent.

5. The computer performance adjustment method based on user facial expression recognition according to claim 4, characterized in that, In step S3, the specific steps of judging the change in the user emotional expression are as follows: S31. regarding each user emotional expression as a random variable, then regarding multiple user emotional expressions as multiple random variables; S32. judging whether two random variables in adjacent identification times are equal, if yes, the user emotional expression remains stable, otherwise, the user emotional expression changes.

6. The computer performance adjustment method based on user facial expression recognition according to claim 5, characterized in that, In step S4, the specific steps of judging the user experience are as follows: S41. Representing a plurality of random variables representing user emotional expressions as an overall quantity for representing a change in user emotional expressions , the overall quantity is expressed in detail as follows: wherein, represents the result of a rank-sum test of happy user emoticon expressions, represents the result of a rank-sum test of contempt user emoticon expressions, represents the result of a rank-sum test of sad user emoticon expressions, represents the result of a rank-sum test of disgust user emoticon expressions, represents the result of a rank-sum test of surprise user emoticon expressions, represents the result of a rank-sum test of anger user emoticon expressions, represents the result of a rank-sum test of fear user emoticon expressions, represents the result of a rank-sum test of indifference user emoticon expressions; S42. Set a threshold for emotional expression changes to judge the overall quantity. If the change in emotional expression exceeds the threshold, the user experience changes; otherwise, the user experience remains stable.

7. The computer performance adjustment method based on user facial expression recognition according to claim 6, characterized in that, In step S41, the user emotional expressions include happy, contempt, sad, disgust, surprise, anger, fear and indifference. If the happy user emotional expression changes, the value of the result of the rank sum test corresponding to the happy user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the happy user emotional expression is 0; if the contempt user emotional expression changes, the value of the result of the rank sum test corresponding to the contempt user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the contempt user emotional expression is 0; if the sad user emotional expression changes, the value of the result of the rank sum test corresponding to the sad user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the sad user emotional expression is 0; if the disgust user emotional expression changes, the value of the result of the rank sum test corresponding to the disgust user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the disgust user emotional expression is 0; if the surprise user emotional expression changes, the value of the result of the rank sum test corresponding to the surprise user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the surprise user emotional expression is 0; if the anger user emotional expression changes, the value of the result of the rank sum test corresponding to the anger user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the anger user emotional expression is 0; if the fear user emotional expression changes, the value of the result of the rank sum test corresponding to the fear user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the fear user emotional expression is 0; if the indifference user emotional expression changes, the value of the result of the rank sum test corresponding to the indifference user emotional expression is 1, otherwise, the value of the result of the rank sum test corresponding to the indifference user emotional expression is 0.

8. A computer performance adjustment system based on user expression recognition, which is implemented based on the computer performance adjustment method based on user expression recognition according to any one of claims 1-7, characterized in that, The system comprises: an acquisition module configured to acquire an expression picture of a user using a computer; an identification module configured to identify the expression picture to obtain a user emotional expression; a first judgment module configured to judge a change in the user emotional expression according to the obtained user emotional expression; a second judgment module configured to judge a user experience by using the change in the user emotional expression; an adjustment module configured to dynamically adjust a computer performance according to the user experience until the computer performance reaches an optimum.

9. The computer performance adjustment system based on user facial expression recognition according to claim 8, wherein, In the identification module, the user emotional expression includes eight user emotional expressions of happy, contempt, sad, disgust, surprise, anger, fear and indifference.

Images