CN109933186B - Mobile web browser energy consumption optimization method based on screen brightness adjustment - Google Patents

Abstract

A mobile Web browser energy consumption optimization method based on screen brightness adjustment comprises the steps of 1) utilizing an interface of an open source project chrome applied by a Webkit to collect data and implement an optimization scheme, compiling the interface to serve as a Web browser tested by an experiment, 2) quantifying impression screen brightness factors, 3) realizing a mobile terminal embedded background system interface, 4) modeling all the factors influencing adjustment factors, processing collected different user data, realizing an SVM classifier, performing cross training on the classifier by using training data stored in a database, finally embedding an optimization model into a mobile terminal, 5) designing a visual compensation scheme after screen brightness adjustment, processing display contents according to screen brightness adjustment factors b and data in a frame cache, and displaying the processed display contents on a mobile terminal screen, wherein the maximum optimization rate is up to 30% by comparing energy consumption in a default screen brightness adjustment mode with energy consumption using an optimization method.

Description

A method for optimizing energy consumption of mobile web browsers based on screen brightness adjustment

technical field

The invention belongs to the technical field of mobile computing, and in particular relates to a method for optimizing energy consumption of a mobile web browser based on screen brightness adjustment.

Background technique

With the continuous innovation of science and technology, smart mobile devices have reached the level of popularity. At the same time, various mobile applications are urgently consuming the power of mobile terminals. The survey found that more than 52% of users use Webkit-based Web browsers every day, and there are many Webki-based applications in addition. According to experimental measurements, the power consumption of screen brightness accounts for a very significant part in the use of mobile phones, but the existing technology does not have a good solution to reduce the energy consumption of mobile terminals by adjusting screen brightness for this part of applications based on large user groups .

Contents of the invention

In order to overcome the deficiencies of the above-mentioned prior art, the object of the present invention is to provide a method for optimizing energy consumption of mobile web browsers based on screen brightness adjustment. This optimization method is based on the screen brightness adjusted by the system according to the light intensity, combined with user experience. Properly reduce the screen brightness to optimize energy consumption. Considering the impact of different web content on user experience, a general model is established for different users using the SVM classification method. By collecting different users, different moments, and different browsing content, the system adjusts the screen brightness according to the light intensity, the screen content and the corresponding users can Accepted screen brightness to train a classification model. When a user uses a Webkit-based application to browse content, analyze in advance the content complexity of the next frame of content to be displayed on the screen in the frame buffer after Webkit rendering and the current screen brightness determined by the light intensity, according to the trained embedded model Dynamically adjust the brightness of the screen, and at the same time perform brightness compensation on the content in the GPUBuffer cache before displaying, which meets the visual needs of users while browsing while ensuring low screen brightness and energy consumption, and reduces the energy consumption of mobile Web applications from the software and hardware level.

In order to achieve the above object, the technical scheme adopted in the present invention is:

A method for optimizing energy consumption of a mobile web browser based on screen brightness adjustment, comprising the following steps:

1) According to the interface of Chromium, an open source project based on Google's Webkit application, the functions include reading the frame buffer data in the GPUBuffer, modifying the frame buffer content according to the brightness adjustment factor, and using it to collect data and implement optimization schemes. After compilation, it will be used as an experimental test web browser;

The open source project Chromium described above is based on the Webkit kernel, supports hardware-accelerated rendering, and has the functions of reading frame buffer content from the browser and GPUBuffer, and writing to GPUBuffer. The openGL context reads the content in the GPUBuffer, which is stored in the RGB space of each pixel of the screen. When using the optimization mode, the value of the brightness-compensated pixel is also written to the GPUBuffer in RGB space;

2) Experiments verify that the main factors affecting user experience when screen brightness changes are browsing content and ambient light intensity, and ambient light intensity directly affects the basic screen brightness adjusted by the system, and quantify these factors;

The selection and quantification of the main factors affecting screen adjustment mentioned above, the light intensity determines the basic screen brightness for system adjustment, on this basis, the screen brightness is appropriately reduced, so the adjusted screen brightness and the system adjusted screen brightness show a positive correlation. , the screen brightness is represented by the screen brightness level value in the mobile phone, and an Android phone is selected for the experiment. In the Android Galaxy S4, the screen brightness level value level is between 1-15, and the screen brightness factor b is calculated for the application level value, b=level /15;

Browsing content On the one hand, the complexity of the content affects the user experience and thus affects the adjustment of the screen brightness. The more complex the content, the higher the user's demand for screen brightness. Here, the entropy of the pixel gray value is used to describe the content complexity; on the other hand, browsing the content also affects the restoration ability after light compensation, thus affecting the adjustable lower limit of the screen brightness. According to experimental statistics, after the screen content is light-compensated, if the brightness value of more than 12% of the pixels cannot be restored, it will affect the user experience. Therefore, 12% is set as the threshold for screen light compensation adjustment. When the pixel When the point ratio exceeds the threshold, the screen brightness adjustment factor obtained by the classifier is linearly increased according to the exceeded ratio

Realize the embedded background system interface of the mobile terminal, the functions include: automatically collect model training data and model prediction data during use, and adjust the screen brightness to implement the optimization plan;

The functions of the background system interface described in step 3) include: automatically collecting model training data and model prediction data during use, and adjusting screen brightness to implement an optimization plan. When a user uses a Webkit-based application to browse, the background system will automatically collect model training data. The data includes: system screen brightness for different users, different times, and different browsing content, screen browsing content, and corresponding user-acceptable screen brightness adjustments factor;

When browsing with the optimized model, the background system will collect user data every 5s. The data includes: the user, the current moment, the current system screen brightness factor, and the browsing content currently stored in the GPU cache as feature values input optimization model;

Modeling is carried out for all the factors that affect the adjustment factor. After processing the collected different user data, design and implement the SVM classifier according to the characteristics of the data samples, and use the training data stored in the database to perform cross-training on the classifier. Finally, the optimized model is embedded in the mobile terminal;

Step 4) Use the training data collected in step 3 as the input value of the SVM optimization model. In order to establish a general model, we calculate the average value of the data of different users and use it as the training data. Use python to set the parameters of the SVM and use the data set Cross-train and generate a classifier according to a 5:1 ratio

Design a visual compensation scheme after screen brightness adjustment, and process the display content according to the screen brightness adjustment factor b and the data in the frame buffer before displaying it on the mobile screen;

Step 5) described visual compensation scheme. After the brightness adjustment of the optimization scheme, the data in the frame buffer will become darker in brightness than the original system adjustment value. In order to make up for this difference and satisfy the user experience, the original RGB image is converted into a YUV image, and then according to The formula Yl=b*min(Y/b,255)≈Y compensates the brightness Y value.

The beneficial effects of the present invention are:

During the use of mobile terminals, screen power consumption is a very significant part. At the same time, in the process of using mobile browsers to browse web pages, users have different screen brightness acceptance for different browsing content, but the current system screen brightness The adjustment mechanism only considers the light intensity, and does not adjust the screen brightness in combination with the user experience, resulting in a waste of energy resources. In view of the above problems, an optimization method is designed to meet the visual needs of users while browsing while ensuring low screen brightness and energy consumption, and reducing the energy consumption of mobile Web applications from the level of software and hardware. The training produced a classifier with up to 98% accuracy.

After testing, when users use two different mobile phones under different environmental lighting and browse different content, comparing the energy consumption in the default screen brightness adjustment mode and the energy consumption using the optimization method, the optimization has a certain optimization ability, and the maximum energy consumption optimization rate as high as 30%.

Description of drawings

Fig. 1 is the work flowchart of the method of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings, but the present invention is not limited to the following embodiments.

A method for optimizing energy consumption of a mobile web browser based on screen brightness adjustment, comprising the following steps:

1) According to the interface of Chromium, an open source project based on Google's Webkit application, the functions include reading the frame buffer data in the GPUBuffer, modifying the frame buffer content according to the brightness adjustment factor, and using it to collect data and implement optimization schemes. After compilation, it will be used as an experimental test web browser;

The open source project Chromium described above is based on the Webkit kernel, supports hardware-accelerated rendering, and has the functions of reading frame buffer content from the browser and GPUBuffer, and writing to GPUBuffer. The openGL context reads the content in the GPUBuffer, which is stored in the RGB space of each pixel of the screen. When using the optimization mode, the value of the brightness-compensated pixel is also written to the GPUBuffer in RGB space;

2) Experiments verify that the main factors affecting user experience when screen brightness changes are browsing content and ambient light intensity, and ambient light intensity directly affects the basic screen brightness adjusted by the system, and quantify these factors;

The selection and quantification of the main factors affecting screen adjustment mentioned above, the light intensity determines the basic screen brightness for system adjustment, on this basis, the screen brightness is appropriately reduced, so the adjusted screen brightness and the system adjusted screen brightness show a positive correlation. , the screen brightness is represented by the screen brightness level value in the mobile phone;

The content complexity of browsing content affects the user experience and thus affects the adjustment of screen brightness. The more complex the content, the higher the user's demand for screen brightness. Therefore, the entropy of the pixel gray value is used to describe the content complexity; browsing content also affects light compensation. The ability to restore in the future affects the adjustable lower limit of the screen brightness. After the screen content undergoes light compensation, if the brightness value of more than 12% of the pixels cannot be restored, it will affect the user experience. Therefore, set 12% As the threshold for screen light compensation adjustment, when the pixel ratio exceeds the threshold, the screen brightness adjustment factor obtained by the classifier is linearly increased according to the exceeded ratio;

3) Realize the interface of the embedded background system of the mobile terminal. The functions of the background system interface include: automatically collecting model training data and model prediction data during use, and adjusting the brightness of the screen to implement an optimization plan; when users use Webkit-based applications to browse, the background system The system will automatically collect model training data, model training data: system screen brightness under different users, different moments, and different browsing content, screen browsing content and corresponding screen brightness adjustment factors acceptable to users; when browsing with the optimized model, The background system will collect user data during use every 5s. The collected data includes: the user, the current moment, the current system screen brightness factor, and the browsing content currently stored in the GPU cache, which are input into the optimization model as feature values;

4) Modeling for all factors that affect the adjustment factor After processing the collected different user data, design and implement the SVM classifier according to the characteristics of the data sample, and use the training data stored in the database to cross the classifier Training, and finally embed the optimization model into the mobile terminal; the collected training data is used as the input value of the SVM optimization model. In order to establish a general model, the average value of the data of different users is calculated as the training data, and the parameters of the SVM are set using python. Use the data set to cross-train and generate a classifier at a ratio of 5:1;

5) Design the visual compensation scheme after the screen brightness adjustment, and display the content on the mobile terminal screen after processing the display content according to the screen brightness adjustment factor b and the data in the frame buffer; the visual compensation scheme, the data in the frame buffer After the brightness adjustment of the optimization scheme, the brightness will become darker compared to the original system adjustment value. The original RGB image is converted into a YUV image, and then according to the formula Yl=b*min(Y/b,255)≈Y pair Brightness Y value is compensated.

Example

1) Invite 20 volunteers to use the developed browser to browse the web, and collect user training data by calling the background interface. Basically, reduce the screen brightness and perform light compensation accordingly, and record the screen brightness adjustment factor acceptable to the user. Take this as a complete training data. For the data when 20 users browse more than 500 web pages, and calculate the average value for the data of different users.

2) Cross-train the SVM classifier according to more than 5000 pieces of data obtained after processing, and embed it into the mobile terminal.

3) During use, the background system collects user data from the browser and the system every 5s, inputs the model, and obtains the model output to the screen brightness adjustment factor b.

4) According to b, make a brightness compensation for the content displayed in the current frame buffer collected, and calculate the proportion of pixels exceeding the light compensation range at the same time. If it exceeds the threshold, perform a linear calculation on the excess part, and adjust the screen brightness adjustment factor b , and finally write the light-compensated content into the frame buffer and then gradually adjust the brightness to the brightness value calculated by b.

Claims (1)

1. A mobile web browser energy consumption optimization method based on screen brightness adjustment is characterized by comprising the following steps:

1) Reading frame cache data in a GPUBuffer according to an interface of an open source project chrome based on Webkit application of Google, modifying frame cache content according to a brightness adjusting factor, acquiring data, implementing an optimization scheme, and compiling the frame cache data to serve as a Web browser for the experimental test;

the open source item chrome supports hardware accelerated rendering based on a Webkit kernel, has the functions of a browser, reading frame cache content from a GPUBuffer and writing the frame cache content into the GPUBuffer, and reads the content in the GPUBuffer through openGL context constructed by the chrome aiming at GPU hardware accelerated rendering in the data acquisition process, wherein the content is stored in an RGB space of each pixel point of a screen. When the optimization mode is used, the value of the pixel point subjected to brightness compensation is written into a GPUBuffer according to the RGB space;

2) The experiment verifies that the main factors influencing the user experience due to the change of the screen brightness comprise browsing content and environment illumination intensity, the environment illumination intensity directly influences the basic screen brightness adjusted by the system, and the factors are quantized;

the main factors influencing the screen adjustment are selected and quantized, the light intensity determines the basic screen brightness adjusted by the system, and the screen brightness is properly reduced on the basis, so that the adjusted screen brightness and the screen brightness adjusted by the system are in a positive correlation relationship, and the screen brightness is represented by a screen brightness grade value in the mobile phone;

the complexity of browsing the content affects the user experience, so that the adjustment of the screen brightness is affected, and the more complex the content is, the higher the requirement of the user on the screen brightness is, so that the content complexity is described by the entropy of the gray value of the pixel point; the browsing content also affects the reduction capability after optical compensation, so that the lower limit of adjustable screen brightness is affected, after the screen content is subjected to optical compensation, if the brightness values of more than 12% of the pixels cannot be reduced, the use experience of a user is affected, so that 12% of the pixels are set as the threshold value of the screen optical compensation adjustment, and when the proportion of the pixels exceeds the threshold value, the screen brightness adjustment factor obtained by the classifier is linearly adjusted upwards according to the exceeding proportion;

3) The method realizes the mobile terminal embedded background system interface, and the background system interface function comprises the following steps: automatically collecting model training data and model prediction data during use, and adjusting screen brightness to implement an optimization scheme; when a user browses by using the Webkit-based application, a background system automatically acquires model training data, wherein the model training data comprises: different users, different moments, system screen brightness under different browsing contents, screen browsing contents and corresponding screen brightness adjustment factors acceptable by the users; when the optimization model is applied to browse, the background system collects user data every 5s during use, and the collected data comprises the following data: the user, the current moment, the current system screen brightness factor and the browsing content currently stored in the GPU cache are used as characteristic values to be input into the optimization model;

4) Modeling aiming at all factors influencing the adjusting factors, processing the acquired different user data, designing and realizing an SVM classifier according to the characteristics of a data sample, performing cross training on the classifier by using training data stored in a database, and finally embedding an optimization model into a mobile terminal; the method comprises the steps that collected training data serve as input values of an SVM optimization model, in order to establish a general model, average values of data of different users are calculated and then serve as training data, python is used for setting parameters of the SVM, then cross training is conducted by utilizing a data set according to the proportion of 5;

5) Designing a visual compensation scheme after the screen brightness is adjusted, processing the display content according to the screen brightness adjustment factor b and the data in the frame buffer, and then displaying the processed display content on the mobile terminal screen; in the visual compensation scheme, after the brightness of the data in the frame buffer is adjusted according to the optimization scheme, the data becomes darker in brightness relative to an original system adjustment value, the original RGB image is converted into a YUV image, and then the brightness Y value is compensated according to a formula of Yl = b × min (Y/b, 255) approximately matching Y;

and processing the display content according to the screen brightness adjusting factor b and the data in the frame buffer, displaying the processed display content on the mobile terminal screen, and comparing the energy consumption in the default screen brightness adjusting mode with the energy consumption of the optimization method, wherein the maximum optimization rate is up to 30%.

Images