WO2020233324A1

WO2020233324A1 - Method and apparatus for testing performance of terminal device

Info

Publication number: WO2020233324A1
Application number: PCT/CN2020/086172
Authority: WO
Inventors: 张起
Original assignee: 深圳壹账通智能科技有限公司
Priority date: 2019-05-21
Filing date: 2020-04-22
Publication date: 2020-11-26
Also published as: CN110232011A; CN110232011B

Abstract

Disclosed in embodiments of the present application are a method and apparatus for testing performance of a terminal device, applicable to generation of a test case. The method comprises: obtaining a central processing unit idle frequency of a terminal device to be tested; obtaining M feature pictures in a test video, and determining the maximum feature picture in the M feature pictures and the maximum feature picture size, the maximum feature picture being a feature picture having the maximum size in the M feature pictures; extracting RGB color values respectively corresponding to pixel points forming the maximum feature picture, and determining a color gamut value of the maximum feature picture according to the RGB color values; and calculating a performance test value according to the maximum feature picture size, the color gamut value, and the central processing unit idle frequency, and determining performance of the terminal device according to the performance test value. By using the embodiments of the present application, the performance of a terminal device can be tested, the performance test mode is diversified, the operation flexibility of performance test is improved, and the applicability is high.

Description

Method and device for testing terminal equipment performance

This application claims the priority of the Chinese patent application filed with the Chinese Patent Office on May 21, 2019, the application number is 201910421940.0, and the invention title is "Method and Apparatus for Testing Terminal Equipment Performance", the entire content of which is incorporated into this application by reference in.

Technical field

This application relates to the field of computer technology, and in particular to a method and device for performance testing of terminal equipment.

Background technique

With the continuous development of network technology, it is no longer as usual for some institutions to conduct business on-site, but can complete business processing operations through remote video between customers and customer service. Among them, the extraction of image features in videos is the prerequisite of the current advanced pattern recognition and artificial intelligence systems. Therefore, feature extraction algorithms applied to terminal equipment can protect the security of loan business through remote video. The inventor realizes that the system performance of the terminal device directly determines the running speed and use efficiency of the video feature extraction algorithm. Therefore, the performance detection of the terminal device is particularly important.

Summary of the invention

The embodiments of the present application provide a method and device for performance testing of terminal equipment. The performance of terminal equipment can be tested, which increases the diversity of performance testing methods, improves the operational flexibility of performance testing, and has high applicability.

In the first aspect, an embodiment of the present application provides a method for testing the performance of a terminal device. The method includes:

Acquiring an idle frequency of the central processing unit of the terminal device to be tested, where the above-mentioned idle frequency of the central processing unit is an idle frequency used to perform image feature extraction;

Obtain M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size among the above M feature pictures, where M is an integer greater than 0, and the above largest feature picture is the largest size among the above M feature pictures Feature picture of

Extracting the RGB color values corresponding to each pixel point constituting the largest feature picture, and determining the color gamut value of the largest feature picture according to each RGB color value;

The performance test value is obtained by calculating the maximum feature image size, the color gamut value, and the idle frequency of the central processing unit, and the performance of the terminal device is determined according to the performance test value.

The embodiment of the application obtains the maximum feature picture size, the color gamut value of the largest feature picture, and the CPU idle frequency in the test video, and then uses the determined maximum feature picture size, color gamut value, and CPU idle frequency to measure The performance of terminal equipment increases the diversity of performance testing methods, improves the operational flexibility of performance testing, and has high applicability.

In the second aspect, an embodiment of the present application provides a terminal device performance test apparatus, which includes:

The frequency acquisition module is used to acquire the idle frequency of the central processor of the terminal device to be tested, and the above-mentioned idle frequency of the central processor is the idle frequency for performing image feature extraction;

The picture acquisition module is used to acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size in the above M feature pictures, where M is an integer greater than 0, and the largest feature picture is the above M The feature picture with the largest size among the feature pictures;

A data determination module for extracting the RGB color values corresponding to each pixel point constituting the largest feature picture, and determining the color gamut value of the largest feature picture according to each RGB color value;

The performance determining module is used to calculate the maximum characteristic picture size, the color gamut value and the idle frequency of the central processing unit to obtain the performance test value, and determine the performance of the terminal device according to the performance test value.

With reference to the second aspect, in a possible implementation manner, the above-mentioned frequency acquisition module is used for:

The maximum CPU frequency and the CPU turbo frequency are acquired, and the CPU frequency difference between the above-mentioned CPU maximum frequency and the above-mentioned CPU turbo frequency is determined as the CPU idle frequency.

In a third aspect, an embodiment of the present application provides a terminal device. The terminal device includes a processor and a memory, and the processor and the memory are connected to each other. The memory is used to store a computer program that supports the terminal device to execute the method provided in the first aspect and/or any one of the possible implementations of the first aspect, the computer program includes program instructions, and the processor is configured to call the foregoing The program instructions execute the method provided in the first aspect and/or any possible implementation manner of the first aspect.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium that stores a computer program, and the computer program includes program instructions that, when executed by a processor, cause the processor to execute The method provided by the foregoing first aspect and/or any possible implementation manner of the first aspect.

The implementation of the embodiments of this application has the following beneficial effects:

The embodiment of the present application determines the performance of the terminal device by obtaining the maximum feature picture size, the color gamut value, and the idle frequency of the central processing unit, which improves the accuracy of the performance test of the terminal device and has high applicability.

Description of the drawings

FIG. 1 is a schematic flowchart of a method for testing performance of a terminal device provided by an embodiment of the present application;

FIG. 2 is a schematic diagram of an application scenario of a terminal device performance testing method provided by an embodiment of the present application;

FIG. 3 is a schematic structural diagram of a terminal equipment performance testing apparatus provided by an embodiment of the present application;

Fig. 4 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.

Detailed ways

The method for testing the performance of terminal devices provided by the embodiments of the present application can be widely applied to the performance testing of terminal devices capable of image recognition. Among them, terminal devices include desktop computers, notebook computers, tablet computers, smart phones, etc., for ease of description, Unified description as terminal equipment. After acquiring the test video, the terminal device can extract multiple frames of pictures from the test video according to the preset extraction time, and perform feature extraction processing on each frame of the extracted multiple frames of pictures to obtain multiple feature pictures. The features in each frame of pictures include user facial features and/or background features. For the convenience of description, the extracted user facial feature pictures and background feature pictures are collectively referred to as feature pictures. Since the size of the feature picture in the convolution calculation of the convolutional neural network will affect the size of the converted image matrix, the embodiment of the present application determines the largest feature picture among multiple feature pictures to measure the performance of the terminal device. Then by processing the color values of the three primary colors (Red, Green, Blue abbreviation, RGB) corresponding to each pixel in the largest feature image, the color gamut value of the largest feature image can be obtained, and then according to the maximum feature image size and color gamut value With the obtained central processing unit (CPU) idle frequency, the performance test value to measure the performance of the terminal device can be determined. By comparing the performance test value with the preset performance test threshold, if the obtained performance test value is greater than or equal to The preset performance test threshold indicates that the performance of the terminal device does not meet the requirements. By adopting the embodiments of the present application, the performance of terminal equipment can be tested, the diversity of performance testing methods is increased, the operational flexibility of performance testing is improved, and the applicability is high.

The methods and related devices provided in the embodiments of the present application will be described in detail below with reference to FIGS. 1 to 4. The method provided in the embodiments of the application may include methods for obtaining CPU idle frequency, obtaining M feature pictures in the test video and determining the largest feature picture, extracting the RGB color value corresponding to each pixel of the largest feature picture, and based on RGB color. The value determines the color gamut value of the above-mentioned largest feature picture and determines the performance of the terminal device and other data processing stages. Among them, the implementation of the above-mentioned data processing stages can be referred to the implementation shown in Figure 1 below.

Refer to FIG. 1, which is a schematic flowchart of a method for testing the performance of a terminal device according to an embodiment of the application. The method provided in the embodiment of the present application may include the following steps 101 to 104:

101. Obtain the idle frequency of the central processing unit of the terminal device to be tested.

In some feasible implementations, Turbo Frequency means that when an application is opened or started, the processor will automatically accelerate to the appropriate frequency, and the original operating speed will be increased by 10% to 20% to ensure smooth running of the application A technology. It is not difficult to understand that there is still a distance between the CPU turbo frequency of the terminal device when the application is in the turbo frequency state and the maximum CPU frequency of the terminal device. In the embodiment of the application, the video review system may be a web video review webpage developed based on a browser, and the feature extraction algorithm used for image feature extraction processing may be a browser plug-in, browser plug-in and application packaged into Programs are used to implement a specific function. The difference is that browser plug-ins depend on the browser and applications rely on terminal devices. Therefore, we can understand browser plug-ins as a special application. Therefore, the CPU turbo frequency can be obtained when the terminal device opens the web video review webpage through the browser, but has not yet used the browser plug-in to unframe to obtain the picture and perform the picture feature extraction process. Optionally, the video review system can also be a client or application. The CPU turbo frequency is obtained when the client or application for video review is opened, but the picture has not been unframed to obtain the picture and the picture feature extraction process is performed. . By obtaining the maximum CPU frequency, and subtracting the maximum CPU frequency from the CPU turbo frequency, the CPU frequency difference between the CPU maximum frequency and the CPU turbo frequency can be obtained. The CPU frequency difference is the CPU idle frequency. In other words, in this embodiment of the present application, the CPU idle frequency refers to the frequency that can be used to perform image feature extraction processing.

102. Obtain M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size among the M feature pictures.

In some feasible implementations, the test video can be the process of interactive operation between the simulated customer and the remote video of the customer service, or the process of real-time interactive operation between the customer and the remote video of the customer service under the real business scene captured by the camera, for example Remote video of customers and customer service to complete the loan review process. In the embodiment of the present application, the test video is a pre-shooting process of simulating the interaction between the customer and the remote video of the customer service, and is a test video stored in the terminal device or in an external storage. The terminal device under test can avoid the problem of inaccurate performance of the terminal device under test due to different video characteristics by reading the same piece of test video for processing. It is understandable that the video is actually played continuously one frame after another. Due to the special physiological structure of the human eye, if the frame rate of the picture seen by the human eye is higher than 16, it will be regarded as coherent. , This phenomenon is called visual stay. This is why the movie film is shot frame by frame and then played quickly. In the embodiment of this application, it can be set to extract N frames of pictures in the test video for feature extraction processing when the test video progresses to a certain point in time. Generally speaking, because the test video shot in the real scene is from the beginning to the stable It takes a period of transition, so N frames of pictures can be extracted for processing when the test video reaches 1 minute. It is not difficult to understand that the features extracted by the feature extraction process include user facial features and/or background features, where the user’s facial features include the user’s eyes, nose, mouth, etc., and the background features include surrounding environment features. In order to facilitate description, the extracted user facial feature pictures and background feature pictures are collectively referred to as feature pictures. By performing feature extraction processing on each of the N frames of pictures separately, M feature pictures in the N frames of pictures can be obtained, where the M feature pictures are a set of feature pictures extracted from the N frames of pictures. Optionally, when determining the maximum feature picture and the maximum feature picture size in M feature pictures, first obtain the feature picture size corresponding to each feature picture in the M feature pictures, and then sort the size of each feature picture in ascending order or Sort in descending order, and finally determine the largest feature picture size and the largest feature picture corresponding to the largest feature picture size based on the sorting result.

In some feasible implementation manners, the most resource-consuming link in the process of facial feature extraction is the process of using convolutional neural networks to perform convolution calculations. Among them, the main calculation method of the convolutional neural network algorithm is to convert the feature picture into an image matrix and then perform operations with a fixed matrix (convolution kernel). It is not difficult to understand that the larger the feature picture, the larger the image matrix converted from the feature picture, so the size of the feature picture will increase exponentially and affect the amount of calculation of convolution. In the embodiment of the present application, after determining the maximum feature picture and the maximum feature picture size from M feature pictures, the maximum feature picture and the maximum feature picture size can be used to measure the performance of the terminal device.

103. Extract the RGB color values corresponding to each pixel point constituting the largest feature picture, and determine the color gamut value of the largest feature picture according to each RGB color value.

In some feasible implementations, since the feature picture is a color picture, the image matrix converted from the feature picture will involve the RGB color value of each pixel. If the feature picture has very rich colors, then the feature picture is converted into an image The result of convolution calculation after matrix is more discrete. Therefore, the RGB color values corresponding to all the pixels constituting the largest feature picture can be extracted, and the color gamut value of the largest feature picture can be determined according to each RGB color value. Among them, the RGB color value is the hexadecimal RGB color value, so each RGB color value can be converted from the hexadecimal RGB color value to the decimal RGB color value to obtain the decimal RGB color value corresponding to each RGB color value. , And then calculate the average value of the decimal RGB color value of each decimal RGB color value, by converting the average value of the decimal RGB color value into a binary RGB color value, the word length of the binary RGB color value can be determined, and the word length is determined as the largest feature picture The color gamut value. Optionally, the hexadecimal average value of each RGB color value can be directly calculated to obtain the hexadecimal RGB color value average value, and the hexadecimal RGB color value average value can be converted into binary RGB color value. The word length of the binary RGB color value can be determined, and then the word length can be determined as the color gamut value of the largest feature picture.

Optionally, in some feasible implementation manners, the RGB color values corresponding to all the pixels that make up the largest feature picture can also be extracted and arranged in ascending or descending order, and then the sorted RGB color values are extracted. The median, after converting the median into a binary median, the word length of the binary median can be determined as the color gamut value of the largest feature picture.

For example, refer to FIG. 2, which is a schematic diagram of an application scenario of the terminal device performance testing method provided by an embodiment of the present application. Suppose that when the test video progresses to one minute, 5 frames of pictures in the test video are extracted for feature extraction processing, and each frame of pictures assumes that 4 feature pictures are extracted, then there are a total of 20 feature pictures in 5 frames of pictures, from Among these 20 feature pictures, the largest feature picture among the 20 feature pictures can be determined. Here, the largest feature picture refers to the feature picture with the largest picture size among the 20 feature pictures. Since the feature image is a color image and the convolution calculation is the color value multiplication after the feature image is converted into an image matrix, the larger the feature image, the more resources it takes, and the richer the color of the feature image, it performs convolution calculation The values obtained later are also more discrete. Therefore, in the embodiment of the present application, the largest feature picture is used to measure the resource of the terminal device occupied by the convolution calculation, and then the performance of the terminal device is measured. Assume that the largest feature picture is the feature picture 1-1 in the first frame of picture, and the feature picture 1-1 includes 4 pixels, where the RGB color values of the 4 pixels are 4169E1, 292421, COC0C0, and E3CF57, respectively. Converting the RGB color values of the above four pixels from the hexadecimal RGB color values to the decimal RGB color values can obtain the respective decimal RGB color values corresponding to each RGB color value of 4289945, 2696225, 12632256, and 14929751. Calculate the average value of the decimal RGB color value of each decimal RGB color value, the average value of the decimal RGB color value can be obtained as 8636294, and the average value of the decimal RGB color value 8636294 is converted into the binary RGB color value, and the binary RGB color value can be obtained as 100000111100011110000110, where , The word length of the binary RGB color value is 24, that is, the color gamut value of the largest feature image is 24.

For another example, suppose that when the test video reaches one minute, 5 frames of pictures in the test video are extracted for feature extraction processing, and each frame of pictures assumes that 4 feature pictures are extracted, so there are a total of 20 pictures in 5 frames. Feature pictures, from these 20 feature pictures, the largest feature picture in the 20 feature pictures can be determined. Here, the largest feature picture refers to the feature picture with the largest size among the 20 feature pictures. Since the feature image is a color image and the convolution calculation is the color value multiplication after the feature image is converted into an image matrix, the larger the feature image, the more resources it takes, and the more the color of the feature image, the more the calculated value Also more discrete. Therefore, in the embodiment of the present application, the largest feature picture may be used to measure the terminal device resources occupied by the convolution calculation, and then to measure the performance of the terminal device. Assume that the largest feature picture is the feature picture 1-1 in the first frame of picture, and the feature picture 1-1 includes 4 pixels, where the RGB color values of the 4 pixels are 4169E1, 292421, COC0C0, and E3CF57, respectively. Converting the RGB color values of the above four pixels from the hexadecimal RGB color values to the decimal RGB color values can obtain the decimal RGB color values corresponding to each RGB color value, namely 4289945, 2696225, 12632256 and 14929751. Extract the RGB color values corresponding to all the pixels that make up the largest feature image and arrange them in ascending order to get the arrangement order as 2696225, 4289945, 12632256, 14929751, and extract the median of each RGB color value after sorting to get the median The number is 8459600, and 100000010001010101010000 can be obtained by converting the above-mentioned median into a binary median, where the word length of the binary median is 24, that is, the color gamut value of the largest feature picture is 24.

104. Calculate the aforementioned maximum characteristic picture size, the aforementioned color gamut value, and the aforementioned CPU idle frequency to obtain a performance test value, and determine the performance of the aforementioned terminal device according to the aforementioned performance test value.

In some feasible implementation manners, the product of the i-th power of the maximum feature picture size and the color gamut value is determined, and then the quotient of the product and the idle frequency of the central processing unit is determined as the performance test value of the terminal device. If the performance test value is less than the preset performance test threshold, it is determined that the performance of the terminal device meets the requirements, where i is an integer greater than 1. In the embodiment of the present application, the performance test value may be the value calculated by multiplying the maximum feature image size by the color gamut value and dividing by the CPU idle frequency. If the performance test value is greater than or equal to the preset performance test threshold, it may be Determine that the performance of the terminal equipment does not meet the requirements.

Optionally, in some feasible implementation manners, after receiving the test video, the terminal device can extract multiple frames of pictures from the test video according to a preset extraction time, and then can perform feature extraction on each frame of the multiple frames of pictures Processing, wherein the features extracted according to the feature extraction algorithm include extracting user facial features and/or background features in each frame of picture. It is not difficult to understand that the processing of the test video will occupy the hardware resources of the terminal device. Among them, the hardware resources of the terminal device include various performance parameters. Among them, various performance parameters include but are not limited to CPU occupancy rate, memory occupancy rate, The graphics processor (Graphics Processing Unit, GPU) occupancy rate, etc., can be specifically determined according to actual application scenarios, and there is no restriction here. By comparing the values of the acquired performance parameters with the thresholds of the corresponding performance parameters, if the CPU occupancy rate, memory occupancy rate and/or graphics processor GPU occupancy rate are respectively greater than or equal to the corresponding CPU occupancy threshold, The memory occupancy threshold and/or the graphics processor GPU occupancy threshold can determine that the performance of the terminal device does not meet the requirements.

In the embodiment of the present application, after the terminal device obtains the test video, it can extract multiple frames of pictures from the test video according to the preset extraction time, and perform feature extraction processing on each of the extracted multiple frames to obtain multiple images. Feature pictures, where the extracted features in each frame of pictures include user facial features and/or background features. For ease of description, the extracted user facial feature pictures and background feature pictures are collectively referred to as feature pictures. Since the size of the feature picture in the convolution calculation of the convolutional neural network will affect the size of the converted image matrix, the embodiment of the present application determines the largest feature picture among multiple feature pictures to measure the performance of the terminal device. Then by processing the RGB color value corresponding to each pixel in the largest feature picture, the color gamut value of the largest feature picture can be obtained, and then according to the maximum feature picture size, color gamut value and the obtained CPU idle frequency, the terminal device can be determined to be measured The performance test value of performance is compared with the preset performance test threshold. If the obtained performance test value is greater than or equal to the preset performance test threshold, it indicates that the performance of the terminal device does not meet the requirements. By adopting the embodiments of the present application, the performance of terminal equipment can be tested, the diversity of performance testing methods is increased, the operational flexibility of performance testing is improved, and the applicability is high.

Refer to FIG. 3, which is a schematic structural diagram of a terminal equipment performance testing apparatus provided by an embodiment of the present application. The terminal equipment performance test apparatus provided by the embodiment of the present application includes:

The frequency acquisition module 31 is configured to acquire the CPU idle frequency of the terminal device to be tested, and the CPU idle frequency is the idle frequency used to perform image feature extraction;

The picture acquisition module 32 is used to acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size among the M feature pictures, where M is an integer greater than 0, and the largest feature picture is the above The feature picture with the largest size among the M feature pictures;

The data determining module 33 is configured to extract the RGB color values corresponding to the respective pixels constituting the largest feature picture, and determine the color gamut value of the largest feature picture according to each RGB color value;

The performance determining module 34 is configured to calculate the maximum feature image size, the color gamut value, and the CPU idle frequency to obtain a performance test value, and determine the performance of the terminal device according to the performance test value.

In some feasible implementation manners, the aforementioned frequency acquisition module 31 is used to:

Obtain the maximum CPU frequency and the CPU core frequency, and determine the CPU frequency difference between the above CPU maximum frequency and the above CPU core frequency as the CPU idle frequency.

In some feasible implementation manners, the above-mentioned picture acquisition module 32 is used to:

Extract N frames of pictures from the test video, and perform feature extraction processing on each of the above N frames of pictures to obtain M feature pictures in the above N frames of pictures, where the M feature pictures are from the above N frames of pictures A collection of facial feature pictures and/or background feature pictures extracted from.

In some feasible implementation manners, the aforementioned RGB color values are hexadecimal RGB color values; the aforementioned data determining module 33 is used for:

Convert each RGB color value from a hexadecimal RGB color value to a decimal RGB color value to obtain each decimal RGB color value corresponding to each of the above-mentioned RGB color values, and calculate the average value of the decimal RGB color value of each of the above-mentioned decimal RGB color values ；

The average value of the decimal RGB color value is converted into a binary RGB color value, the word length of the binary RGB color value is determined, and the word length is determined as the color gamut value.

In some feasible implementation manners, the aforementioned data determination module 33 is used to:

Converting each RGB color value from a hexadecimal RGB color value into a decimal RGB color value to obtain each decimal RGB color value corresponding to each of the above-mentioned RGB color values;

Extracting the median of the RGB color values in each of the aforementioned decimal RGB color values, and converting the median of the aforementioned RGB color values into binary RGB color values;

The word length of the binary RGB color value is determined, and the word length is determined as the color gamut value.

In some feasible implementation manners, the aforementioned performance determining module 34 is used to:

Determine the product of the i-th power of the maximum feature image size and the color gamut value, where i is an integer greater than 1;

Determining the quotient of the product and the idle frequency of the central processor as the performance test value of the terminal device;

If the above performance test value is greater than or equal to the preset performance test threshold, it is determined that the performance of the above terminal device does not meet the requirements.

Acquiring the feature picture size corresponding to each of the above M feature pictures;

Sort each feature picture size, and determine the maximum feature picture size and the maximum feature picture corresponding to the above-mentioned maximum feature picture size based on the sorting result.

In a specific implementation, the above-mentioned terminal equipment performance testing apparatus can execute the implementation manners provided in the above-mentioned steps in FIG. 1 through various built-in functional modules. For example, the aforementioned frequency acquisition module 31 can be used to perform the implementation methods of acquiring the CPU core frequency, acquiring the maximum CPU frequency, and determining the idle frequency of the CPU in the aforementioned steps. For details, please refer to the implementation methods provided in the aforementioned steps, and will not be repeated here. . The above-mentioned picture acquisition module 32 can be used to perform the implementation described in the relevant steps of extracting pictures in the test video, extracting feature pictures in the pictures, and determining the largest feature picture in the above steps. For details, please refer to the implementation manners provided in the above steps. , I won’t repeat it here. The above-mentioned data determination module 33 can be used to perform implementation methods such as extracting the RGB color value of the largest feature picture in the above steps and determining the color gamut value based on each RGB color value. For details, please refer to the implementation methods provided in the above steps. Repeat. The above-mentioned performance determining module 34 can be used to perform the above-mentioned various steps in the implementation of determining the performance of the terminal device based on the maximum feature picture size, color gamut value, and CPU idle frequency. For details, please refer to the implementation methods provided in the above-mentioned steps, which will not be repeated here. .

In the embodiment of the present application, the terminal device performance test apparatus can be based on the acquired test video, and can extract multiple frames of pictures from the test video according to a preset extraction time, and perform features on each frame of the extracted multiple frames of pictures The extraction process can obtain multiple feature pictures, where the extracted features in each frame of pictures include user facial features and/or background features. For the convenience of description, the extracted user facial feature pictures and background feature pictures are collectively referred to as feature pictures. Since the size of the feature picture in the convolution calculation of the convolutional neural network will affect the size of the converted image matrix, the embodiment of the present application determines the largest feature picture among multiple feature pictures to measure the performance of the terminal device. Then by processing the RGB color value corresponding to each pixel in the largest feature picture, the color gamut value of the largest feature picture can be obtained, and then according to the maximum feature picture size, color gamut value and the obtained CPU idle frequency, the terminal device can be determined to be measured The performance test value of performance is compared with the preset performance test threshold. If the obtained performance test value is greater than or equal to the preset performance test threshold, it indicates that the performance of the terminal device does not meet the requirements. By adopting the embodiments of the present application, the performance of the terminal equipment can be tested, the diversity of performance testing methods is increased, the accuracy of the performance testing of the terminal equipment is improved, the flexibility is high, and the application range is wide.

Refer to FIG. 4, which is a schematic structural diagram of a terminal device provided by an embodiment of the present application. As shown in FIG. 4, the terminal device in this embodiment may include: one or more processors 401 and a memory 402. The aforementioned processor 401 and memory 402 are connected through a bus 403. The memory 402 is configured to store a computer program, and the computer program includes program instructions. The processor 401 is configured to execute the program instructions stored in the memory 402, and perform the following operations:

Obtain the CPU idle frequency of the terminal device to be tested, where the CPU idle frequency is the idle frequency used to perform image feature extraction;

The performance test value is obtained by calculating the maximum feature picture size, the color gamut value, and the CPU idle frequency, and the performance of the terminal device is determined according to the performance test value.

In some feasible implementation manners, the aforementioned processor 401 is configured to:

In some feasible implementation manners, the aforementioned RGB color value is a hexadecimal RGB color value; the aforementioned processor 401 is configured to:

Convert each RGB color value from a hexadecimal RGB color value to a decimal RGB color value to obtain each decimal RGB color value corresponding to each of the above-mentioned RGB color values;

If the performance test value is greater than or equal to the preset performance test threshold, it is determined that the performance of the terminal device does not meet the requirements.

It should be understood that in some feasible implementation manners, the aforementioned processor 401 may be a central processing unit (CPU), and the processor may also be other general-purpose processors or digital signal processors (DSP). , Application specific integrated circuit (ASIC), ready-made programmable gate array (field programmable gate array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like. The memory 402 may include a read-only memory and a random access memory, and provides instructions and data to the processor 401. A part of the memory 402 may also include a non-volatile random access memory. For example, the memory 402 may also store device type information.

In specific implementation, the above-mentioned terminal device can execute the implementation manners provided in each step in FIG. 1 through its built-in functional modules. For details, refer to the implementation manners provided in the foregoing steps, and details are not described herein again.

In this embodiment of the application, the terminal device can extract multiple frames of pictures from the test video based on the acquired test video according to a preset extraction time, and perform feature extraction processing on each of the extracted multiple frames of pictures to obtain Multiple feature pictures, where the features in each frame of the extracted picture include user facial features and/or background features. For ease of description, the extracted user facial feature pictures and background feature pictures are collectively referred to as feature pictures. Since the size of the feature picture in the convolution calculation of the convolutional neural network will affect the size of the converted image matrix, the embodiment of the present application determines the largest feature picture among multiple feature pictures to measure the performance of the terminal device. Then by processing the RGB color value corresponding to each pixel in the largest feature picture, the color gamut value of the largest feature picture can be obtained, and then according to the maximum feature picture size, color gamut value and the obtained CPU idle frequency, the terminal device can be determined to be measured The performance test value of performance is compared with the preset performance test threshold. If the obtained performance test value is greater than or equal to the preset performance test threshold, it indicates that the performance of the terminal device does not meet the requirements. By adopting the embodiments of the present application, the performance of the terminal equipment can be tested, the diversity of performance testing methods is increased, the accuracy of the performance testing of the terminal equipment is improved, the flexibility is high, and the application range is wide.

The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium may be nonvolatile or volatile. The computer-readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by the processor, the method for performing the performance test of the terminal device provided in each step in FIG. 1 can be referred to for details of the implementation manner provided in each step above, which will not be repeated here.

The foregoing computer-readable storage medium may be the terminal device performance testing apparatus provided in any of the foregoing embodiments or the internal storage unit of the foregoing terminal device, such as a hard disk or memory of an electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), or a secure digital (SD) card equipped on the electronic device. Flash card, etc. Further, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

Claims

A method for terminal equipment performance testing, wherein the method includes:

Acquiring a central processor idle frequency of the terminal device to be tested, where the central processor idle frequency is an idle frequency used to perform image feature extraction;

Acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size in the M feature pictures, where M is an integer greater than 0, and the largest feature picture is the M feature pictures The largest feature picture in the size;

Extracting the RGB color values corresponding to each pixel point constituting the largest feature picture, and determining the color gamut value of the largest feature picture according to each RGB color value;

The maximum feature picture size, the color gamut value and the idle frequency of the central processing unit are calculated to obtain a performance test value, and the performance of the terminal device is determined according to the performance test value.
The method according to claim 1, wherein said obtaining the idle frequency of the central processing unit of the terminal device to be tested comprises:

Obtain the maximum CPU frequency and the CPU turbo frequency, and determine the CPU frequency difference between the CPU maximum frequency and the CPU turbo frequency as the CPU idle frequency.
The method according to claim 1 or 2, wherein said obtaining M feature pictures in the test video comprises:

Extract N frames of pictures in the test video, and perform picture feature extraction processing on each frame of the N frames of pictures to obtain M feature pictures in the N frames of pictures, where the M feature pictures are from A collection of facial feature pictures and/or background feature pictures extracted from the N frames of pictures.
The method according to any one of claims 1 to 3, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest characteristic picture according to each RGB color value comprises:

Convert each RGB color value from a hexadecimal RGB color value to a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value, and calculate the decimal RGB color value of each decimal RGB color value average value;

The average value of the decimal RGB color value is converted into a binary RGB color value, the word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The method according to any one of claims 1 to 3, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest characteristic picture according to each RGB color value comprises:

Converting each RGB color value from a hexadecimal RGB color value into a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value;

Extracting the median of the RGB color values in each of the decimal RGB color values, and converting the median of the RGB color values into binary RGB color values;

The word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The method according to claim 4 or 5, wherein the performance test value is obtained by calculating the maximum feature picture size, the color gamut value and the idle frequency of the CPU, and the performance test value is determined according to the performance test value The performance of the terminal equipment includes:

Determining the product of the i-th power of the maximum feature picture size and the color gamut value, where i is an integer greater than 1;

Determining the quotient of the product and the idle frequency of the central processing unit as the performance test value of the terminal device;

If the performance test value is greater than or equal to the preset performance test threshold, it is determined that the performance of the terminal device does not meet the requirements.
The method according to any one of claims 1 to 6, wherein said determining the largest feature picture and the largest feature picture size in the M feature pictures comprises:

Acquiring a feature picture size corresponding to each of the M feature pictures;

Sort each feature picture size, and determine the largest feature picture size and the largest feature picture corresponding to the largest feature picture size based on the sorting result.
A device for performance testing of terminal equipment, wherein the device includes:

A frequency acquisition module, configured to acquire the idle frequency of the central processing unit of the terminal device to be tested, where the idle frequency of the central processing unit is an idle frequency used to perform image feature extraction;

The picture acquisition module is used to acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size in the M feature pictures, where M is an integer greater than 0, and the largest feature picture is The feature picture with the largest size among the M feature pictures;

A data determination module, configured to extract the RGB color values corresponding to each pixel point constituting the largest feature picture, and determine the color gamut value of the largest feature picture according to each RGB color value;

The performance determining module is configured to calculate the maximum characteristic picture size, the color gamut value and the idle frequency of the central processing unit to obtain a performance test value, and determine the performance of the terminal device according to the performance test value.
A terminal device, which includes a processor and a memory, and the processor and the memory are connected to each other;

The memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to perform a method of performing a terminal device performance test, the method of terminal device performance testing including:

Acquiring a central processor idle frequency of the terminal device to be tested, where the central processor idle frequency is an idle frequency used to perform image feature extraction;

Acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size in the M feature pictures, where M is an integer greater than 0, and the largest feature picture is the M feature pictures The largest feature picture in the size;

Extracting the RGB color values corresponding to each pixel point constituting the largest feature picture, and determining the color gamut value of the largest feature picture according to each RGB color value;

The maximum feature picture size, the color gamut value and the idle frequency of the central processing unit are calculated to obtain a performance test value, and the performance of the terminal device is determined according to the performance test value.
The terminal device according to claim 9, wherein said acquiring M feature pictures in the test video comprises:

Extract N frames of pictures in the test video, and perform picture feature extraction processing on each frame of the N frames of pictures to obtain M feature pictures in the N frames of pictures, where the M feature pictures are from A collection of facial feature pictures and/or background feature pictures extracted from the N frames of pictures.
The terminal device according to claim 9 or 10, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest characteristic picture according to each RGB color value includes:

Convert each RGB color value from a hexadecimal RGB color value to a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value, and calculate the decimal RGB color value of each decimal RGB color value average value;

The average value of the decimal RGB color value is converted into a binary RGB color value, the word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The terminal device according to claim 9 or 10, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest characteristic picture according to each RGB color value includes:

Converting each RGB color value from a hexadecimal RGB color value into a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value;

Extracting the median of the RGB color values in each of the decimal RGB color values, and converting the median of the RGB color values into binary RGB color values;

The word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The terminal device according to claim 11 or 12, wherein the performance test value is obtained by calculating the maximum feature picture size, the color gamut value and the idle frequency of the central processing unit, and the performance test value is obtained according to the performance test The value determines the performance of the terminal device, including:

Determining the product of the i-th power of the maximum feature picture size and the color gamut value, where i is an integer greater than 1;

Determining the quotient of the product and the idle frequency of the central processing unit as the performance test value of the terminal device;

If the performance test value is greater than or equal to the preset performance test threshold, it is determined that the performance of the terminal device does not meet the requirements.
The terminal device according to any one of claims 9-13, wherein the determining the largest feature picture and the largest feature picture size in the M feature pictures comprises:

Acquiring a feature picture size corresponding to each of the M feature pictures;

Sort each feature picture size, and determine the largest feature picture size and the largest feature picture corresponding to the largest feature picture size based on the sorting result.
A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program includes program instructions that, when executed by a processor, cause the processor to perform a terminal device performance test The method for testing the performance of the terminal equipment includes the following steps:

Acquiring a central processor idle frequency of the terminal device to be tested, where the central processor idle frequency is an idle frequency used to perform image feature extraction;

Acquire M feature pictures in the test video, and determine the largest feature picture and the largest feature picture size in the M feature pictures, where M is an integer greater than 0, and the largest feature picture is the M feature pictures The largest feature picture in the size;

Extracting the RGB color values corresponding to each pixel point constituting the largest feature picture, and determining the color gamut value of the largest feature picture according to each RGB color value;

The maximum feature picture size, the color gamut value and the idle frequency of the central processing unit are calculated to obtain a performance test value, and the performance of the terminal device is determined according to the performance test value.
The computer-readable storage medium according to claim 15, and the method according to claim 1 or 2, wherein said obtaining M feature pictures in the test video comprises:

Extract N frames of pictures in the test video, and perform picture feature extraction processing on each frame of the N frames of pictures to obtain M feature pictures in the N frames of pictures, where the M feature pictures are from A collection of facial feature pictures and/or background feature pictures extracted from the N frames of pictures.
The computer-readable storage medium according to claim 15 or 16, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest feature picture according to each RGB color value includes :

Convert each RGB color value from a hexadecimal RGB color value to a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value, and calculate the decimal RGB color value of each decimal RGB color value average value;

The average value of the decimal RGB color value is converted into a binary RGB color value, the word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The computer-readable storage medium according to claim 15 or 16, wherein the RGB color value is a hexadecimal RGB color value; the determining the color gamut value of the largest feature picture according to each RGB color value includes :

Converting each RGB color value from a hexadecimal RGB color value into a decimal RGB color value to obtain each decimal RGB color value corresponding to each RGB color value;

Extracting the median of the RGB color values in each of the decimal RGB color values, and converting the median of the RGB color values into binary RGB color values;

The word length of the binary RGB color value is determined, and the word length is determined as a color gamut value.
The computer-readable storage medium according to claim 15 or 16, wherein the calculation of the maximum feature picture size, the color gamut value, and the idle frequency of the CPU obtains a performance test value, and the performance test value The performance test value to determine the performance of the terminal device includes:

Determining the product of the i-th power of the maximum feature picture size and the color gamut value, where i is an integer greater than 1;

Determining the quotient of the product and the idle frequency of the central processing unit as the performance test value of the terminal device;

If the performance test value is greater than or equal to the preset performance test threshold, it is determined that the performance of the terminal device does not meet the requirements.
The computer-readable storage medium according to claim 15 or 16, wherein the determining the largest feature picture and the largest feature picture size in the M feature pictures comprises:

Acquiring a feature picture size corresponding to each of the M feature pictures;

Sort each feature picture size, and determine the largest feature picture size and the largest feature picture corresponding to the largest feature picture size based on the sorting result.