Abstract
Face recognition can convey a kind of intuitive and rich information to people. With the current Internet security problems becoming more and more prominent today, in order to better effectively guarantee the security of such digital images during Internet transmission, one of the most direct and effective solutions is to recognize the characteristics of such digital information on facial expressions. However, the customer service system of the beauty and health club itself is a customer service software system based on customer experience and marketing management as its core functions. This article discusses the main software design and system realization of the management platform of the beauty and health club customer information service system in the entire operation process from the application development trend of information software. Based on the theoretical research and application development of the current corporate customer business service management system of the beauty and health club, the basic concepts and operating methods of current corporate customers, business service information management, and business service customer management, as well as the corresponding corporate information customer management. The management technology and other issues are first briefly summarized in theory, and then based on the specific business development characteristics of the current beauty and health clubs, the main application functions and specific market requirements of the service company’s corporate customer service management system are analyzed in detail and introduced the related functions of using the software.
1. Introduction
The development of mobile Internet information technology has made people’s information work and daily life more convenient, happier, and more colorful [1]. However, in this convenient and colorful mobile Internet era, there are still many hidden network security risks and health in life threatened. Internet technology is an open information network with many insecure factors [2]. When recognizing user faces or expressions or transmitting user signal information, there may still be many opportunities for malicious tampering, sneak attacks, and other illegal copying in violation of regulations. The traditional standard algorithm for group data encryption of facial expressions is not suitable for the facial expression recognition data encryption. The main reason is that these encryption algorithms do not use grouping algorithms that are specifically used for data encryption for the recognition of the internal information of the expression text. The characteristics of large amount of data, high redundancy, uneven distribution of energy information, and high correlation with multiple adjacent encrypted pixels are main encryption features of digital image encryption information [3]. The special nature of these characteristics also leads to the difference between traditional digital grouped image encryption algorithms have extremely different encryption efficiency for the application of grouped encryption of digital images [4]. In the current information security application field, many technical methods have emerged, but the gray value of the image has not changed in any way. The attacker can even use some statistical characteristics such as the gray histogram distribution of facial expressions or exhaustive method to crack these cipher-text images. Therefore, traditional network image information encryption algorithms may be more vulnerable when facing multiple different network attacks at the same time [5]. China Beauty Club, which uses facial expression information recognition, is also a leading enterprise that has developed rapidly in the application field of this technology and is constantly growing. In recent years, with the further expansion of its main business and asset scale, information management has developed into an important system construction content for future enterprise development [6]. This paper will discuss the application software design and hardware realization of the customer service information security management system of beauty and health clubs from the perspective of application software development [7]. The customer service management system of the beauty and health center is a comprehensive software system whose main core function is to provide customer management services. The main purpose is to support enterprise managers to efficiently, comprehensively, and accurately control their company’s various businesses through network technology and to help the company achieve effective management and operation of data [8]. According to the business characteristics of the beauty and health club, the functional requirements of the club’s customer service information management software system are analyzed [9].
2. Related Work
In the early twenty-first century, the literature proposed a multiscale image analysis CT transform based on facial expression recognition. This transform refers to the design and improvement of the traditional existing transform module by adopting the structure of the dual filter module group [10]. It can effectively reflect various geometric changes in facial expressions. CT transformation is composed of two components. The first component is the Laplacian Pyramid, which is mainly responsible for the multiscale decomposition and processing of image signals [11]. The second component refers to the image analysis of the subspace bandwidth after the decomposition of the DLP function in the first part, which can directly select the value of the image directionality through the image directional reflection filter [12]. The literature mainly expresses that the displayed facial recognition image data usually has two-dimensional and distributed characteristics. Generally, we encrypt its image source in the form of scrambling and diffusion [13]. However, this type of facial expression image using encryption diffusion algorithm not only has a relatively low-diffusion efficiency, but it also has many problems due to the complex calculation of diffusion efficiency and low security. Researchers and others proposed the SPCME digital image block encryption algorithm based on the chaotic code replacement structure based on the encryption technology and research of the chaotic program replacement structure of image block cipher [14]. The literature describes in detail that the library management information system is a new type of comprehensive general-purpose software designed, developed, and used specifically for the management, storage, protection, data access, management access, and statistical data processing of the data system [15]. In the design of the library management system, it is widely regarded as an important type of database transaction processing information system used in different data-intensive applications [16]. A database system is an entity that relies on its so-called “relational data simulation” to represent information internally. A relationship is expressed by a set of attribute data. An instance of a relationship is a set of unit tuples with specific attribute values, which become the corresponding relationship scale. As the market operation of contemporary, Chinese enterprises has gradually shifted from the traditional model with product service as the core to the modern model with customer service as the core. Enterprises need to further create a system that can effectively support and cultivate the discovery of customers’ [6] potential and actual needs and their anticipated capabilities, tools, and methods. Taking customers as the service center is to enable companies to accurately identify their expectations and actual needs for customers and to respond quickly to the rapidly evolving customers. Therefore, various industries must transform their marketing management business models from the perspective of customers as the main service body, and the goal is to continuously meet the unique and differentiated needs of each type of customer [17]. The literature shows that the main business object in the customer service management system of the beauty and health club is the customers of the beauty and health club [18]. Therefore, the customer service and management system is an important technical basis for the information management of the beauty and health club [19, 20]. As a software system with unique customer relationship management information system functions and features, the information processing system must store all customer information.
3. Improvement of Facial Expression Recognition Method Based on Image Encryption Algorithm
3.1. Image Encryption Algorithm
Cryptography generally refers to the study of how to conceal or effectively preserve information. There is a close and indivisible relationship between many majors and fields such as cryptography and computer science. This is a discipline with comprehensive value. The generation of ciphers was mainly developed to adapt to the needs of human warfare in the early stage. It consists of two parts: cryptanalysis and cryptography.
With the rapid development of modern information technology, security issues in the field of digitalization and multimedia applications have become increasingly prominent. Video, images, etc., are all important parts of the multimedia application field. It is an image as one of the important ways and means of various multimedia dissemination of information, which can vividly convey the information it wants to display and express to people. In today’s environment with global Internet security problems, in order to effectively ensure the security of digital images, it must be strictly encrypted.
At present, the main types of image encryption technology used in the field of network information encryption technology are the following: first, network information encryption technology based on secret information sharing; second, digital image encryption technology based on matrix transfer transformation; third, based on chaos theory digital image encryption technology. In the encryption processing technology of these images, most of them use scrambling processing technology. Different image encryption technologies have different characteristics of image encryption technology, and the complexity, security, and speed of encryption and decryption of these image encryption technologies are also different.
The definition of Henon mapping is shown in formula (1):.
Henon mapping has a single-valued inverse, and its inverse mapping is shown in formula (2):
Figure 1 shows the phase diagram of Henon mapping.
The security performance of the image encryption algorithm can be detected by the analysis technology of network security. The security requirements are the following: a guaranteed image encryption system must have the characteristics of small correlation between adjacent pixels and balanced distribution of the number of histograms, so as to effectively prevent and control various attacks.
After the pixel value of a certain pixel is cracked, input any number of data information on the neighboring pixels around the pixel:
In formula (3), x and y, respectively, represent the average gray value of each element in an image. Co(x, y) is an organic correlation function representing the threshold of each pixel, E(x) is the average value, and D(x) is the average variance.
represents the probability of the gray level. The formula is shown in equation (4):
For analyzing images, the gray-scale and histogram analysis of the image is a convenient and quick analysis tool. The formula for finding the positive value of the histogram is shown in formula (5), and the formula for finding the variance of the histogram is shown in formula (6).
The calculation of NPCR and UACI can be realized by formulas (7) and formula (8), where M and N represent the height and width of the and images.
In digital image encryption technology, image-scrambling technology is an important part of it. The image scrambling technology can realize the re-encoding and arrangement of each element and pixel matrix in the image, reduce the correlation of each element in the image, and then achieve the goal of image encryption and secure data transmission. In modern mathematics, the number sequence formula in Fibonacci takes a recursive function form for definition, such as formula (9):
Definition of two-dimensional isometric image scrambling transformation:
According to the encryption formula (10), the encryption parameters are set as follows: a = 1, b = 1, c = 1, d = 0. The encryption method obtained is to encrypt the two-dimensional isometric image Fibonacci in the image scrambling after the transformation. As shown in the calculation formula (11):
According to the encryption formula (10), the encryption method obtained by setting the encryption parameters as a = 1, b = 1, c = 1, d = 2 is actually an encryption function transformation of Arnold through the image input scrambler. As shown in the calculation formula (12):
Specific steps of image encryption algorithm based on DNA encoding: The first step: the use of image encoding encryption file mapping needs to choose the logistic mapping processing system. Consider a key mapping as the initial function value of a function as logistic and its mapping. Among them, the method of initially obtaining the key is to select an integer of 16 points from any point between 0 and 255. Such a smart key lock requires large enough space performance and can achieve relatively high security and efficiency in actual operation and use. Step 2: differentiate or calculate each pixel value of the original plaintext image through the initial key. Step 3: take the above two key codes as a key parameter and initialized value of logistic key mapping, respectively, and then replace it with codes again and use it to re-encode to generate a key sequence. Step 4: encode the DNA sequence generated in Step 3 and the value of each element in the original plaintext image separately according to the coding schedule. This sequence is to convert the data information of each sequence and other images into DNA sequences. Step 5: Perform alternative encryption operations on the image through calculation, as shown in Figure 2.
3.2. Improvement of Facial Expression Recognition Method
According to the different regional characteristics of the current facial expression local information feature distribution, this paper mainly proposes a new facial expression information feature recognition and analysis algorithm based on the widened local nearest Gabor neighbor feature and the broadened nearest local neighbor feature classifier. The practice of the model adopted on the JAFFE model database has proved that this method has achieved good results.
The core idea of facial expression recognition based on linear methods is to find a suitable transformation matrix W, and map the original sample x to the feature space y, as shown in formula (13):
The obtained feature vector is classified by the nearest neighbor classifier NN. The sample y category in the test set is the same as the training set sample category in the test set, which satisfies the formula (14):where is the Euclidean distance between the two-dimensional feature vectors Xk and y.
The total divergence matrix A is defined as follows:where is the mean value of the input sample image. The divergence of the feature vector obtained after the transformation of formula (15) is . In PCA, the selection criterion of the transformation matrix W is to maximize the determinant of the total divergence matrix after projection, as shown in formula (16).
Equation (18) can be used to reconstruct the image:
According to the number of principal components m selected, the reconstructable image is shown in Figure 3. The first image on the left is the original image, and the following images are the images reconstructed when 3, 5, 10, 20, and 40 are selected in turn.
The original expression image A usually needs to be converted into a one-dimensional vector B. Some information may be lost in the conversion process, and calculating the total divergence matrix field C is a relatively large workload. 2D-PCA is the expansion of PCA in two-dimensional space, and it can directly map the input image matrix to the feature space.
Assuming that x is an n-dimensional vector and A is a matrix of, which represents the input image, then the projection of A on x can get the feature y:
The quality of the projection vector x can be measured by using the total divergence matrix of the sample after projection. It is easy to know that the total divergence matrix after projection is shown in formula (20):
Define the image divergence matrix as shown in the following formula:
Then formula (21) can be expressed by formula (22):
Figure 4 shows the effect of FLD and PCA on the mapping of two-dimensional samples to the feature space. It can be seen from the figure that the samples after FLD mapping are more separable than PCA.
The main data set used in the experiment is JAFFE, a famous Japanese female expression library. This article will be divided into two types of experiments: “human-dependent facial expression recognition” and “human-independent facial expression recognition.” In “person-independent facial expression recognition,” the strategy of “dropping a sample” is adopted, that is, each female’s facial expressions are taken out as a training sample, and these residual samples are used as test samples. The training set contains 70 samples, and the test set contains 143 samples.
Table 1 shows the highest recognition rate of seven expressions based on the PCA method, and each row represents the probability of a certain expression finally classified into these seven expressions.
Table 2 shows the highest recognition rate of seven expressions based on the 2D-PCA method.
Figure 5 shows the results of facial expression recognition experiments with four linear methods of PCA, 2D-PCA, FLD, and 2D-FLD. The figure shows the variation of recognition rate with feature dimensions.
This step is mainly to cut and segment the expression image, as shown in Figure 6. The expression features are mainly concentrated in the facial area. First, the expression image is cut into a size of 144 × 127. In order to obtain local expression features, the expression image is then divided into areas that overlap each other by 50%. Assuming that the length (width) of the original image is n times the length (width) of the cropped image, then an image can be divided into blocks of local regions. The selection of n will affect the performance and recognition rate. In this paper, n is equal to 4, which means that the original image is divided into 49 local regions.
4. Design and Implementation of Beauty and Health Management System
4.1. Design of Beauty Health Management System
4.1.1. Database ER Model
The beauty and health management system is currently mainly running in the clubhouse, using the client-server model. The background of the local system is used as a general database platform to centralize and manage the business, and it also allows users to access the internal database. Obtain authorized and globally shared information. The attributes and data types of the main entity data of the database are described in Tables 3 and 4.
4.1.2. Rear Terminal Module
An important basic function of the customer service management information system is to achieve quantitative statistics and management for the comprehensive business information and data of customer service. To use this software, we must first obtain the data from the comprehensive business database of our company’s customer service operations. Customer service-related data and complete data processing, and then saving these processing results in the enterprise’s customer service management information system integrated business database to facilitate the account manager to perform statistics and analysis on subsequent customer service data [21]. The main internal structure of the database subsystem is used to perform access and statistical analysis of the integrity of customer data and service information and comprehensive business data and realize the internal data storage process as a database server. These processes are based on the original business comprehensive data records for the various responsibilities of the company and the account manager. The statistical data that need to be collected for query and calculation of business indicators need to be automatically extracted, and all the statistical results are integrated into a complete report for output at the same time, in the case of certain failures or abnormalities in the enterprise, and the information missing due to the failure has been reliably modified and restored. These warehousing procedures are also to realize the statistics and analysis of comprehensive business indicators, backup, dump, and database performance adjustment management of customer information and business data.
4.1.3. Front Terminal Module
The database publishes all the information in the background to the frontend through the Web server and spreads to the backend of users. The system’s role configuration mechanism based on the database platform can complete the allocation of functional roles and access authorization for all legal users. The software in the system regulates the functions and data of which modules are allowed to be accessed based on its configured permissions.
Health care products, because health care products themselves are not only a subsidiary component of various health care and medical beauty services but also one of the main ways for the health club to obtain profit; therefore, we need to effectively target this category in this module. Information is used to achieve a set of professional management responsibilities. The information objects that need to be managed mainly include the names, current quantities, and suppliers of various types of health-care medical products currently in use.
The service item information records the development status of the beauty and health product items for each customer in detail. At the same time, the customer manager also records in detail the physical and mental health of each customer’s feedback or return visit according to his actual situation. The purpose of the key recommendations is to enable the system to accumulate detailed dynamic data that occurred during the process of receiving services from customers within the company, so as to facilitate a comprehensive analysis of specific customers and overall customer satisfaction, potential demand signals, and market trends. At the same time, for all customer managers, it can constitute the most valuable enterprise operation management knowledge within the company.
The service function of the service information record management clubhouse is mainly through the above direct service data analysis and record management methods to indirectly statistics and export the clubhouse business service information related to its project. These business information are the basis for the system to be in the above service project data database, direct calculation and analysis on the above.
4.1.4. Customer Service Information Management
Customer service information management and its statistical analysis subsystem on the one hand provides more in-depth multidimensional statistical data analysis and global business operation data analysis for the customers and service personnel of the club; on the other hand, it provides the company’s group companies with a global operation data and business analysis basis.
4.1.5. Module Composition
According to the analysis of customer service needs, in the company’s customer service information management software, relatively complex large-scale data processing functions require statistical calculations based on various marketing and customer needs analysis. To some extent, these indicators reflect customer satisfaction with actual services.
This system not only calculates the above statistical indicators within the specified time but also calculates the proportions related to the growth and change of various similar indicators in the corresponding year, so as to show the market trend that is valuable to the business manager of the enterprise information. For the management and decision-making level, the information provided by this module directly helps the company to effectively improve the company’s internal operation and management.
This module is based on the cost data information calculated and calculated, and pays special attention to the realization of the internal cost of service projects, which can promote the management and decision-making level to grasp the actual profit space and an important direction for cost risk control.
By analyzing the operating value of a project, this module enables the management to know more clearly which service projects should be profitable and attractive to customers, and what are the main factors that lead to their profitability, which makes the technical personnel in the market more targeted development oriented of products and services with higher profitability.
This module is based on statistics and analysis of customer satisfaction indicators so that account managers and their marketing directors can accurately grasp the actual attractiveness of different types of service items to customers, all factors that attract customer needs, and other influences trends in customer needs.
The function of statistical analysis of operating costs is also realized through a separate set of procedures and links for hierarchical operations. As a service company, the club’s operating cost composition and its measurement and accounting processing intermediate costs are quite large. The refinement and measurement of the company’s operating costs also need to rely to a limited extent on more accurate and complete indirect cost data recording and reasonable cost allocation and calculation rules. The collection and processing of daily cost accounting data are mainly realized by the functions of the respective modules of the system. The processing principle of cost accounting is mainly carried out by the system on the basis of all cost data records within the scope of the system management module. The processing flow is shown in Figure 7.
4.2. Realization of Beauty and Health Management System
4.2.1. Program Development Project Organization
The goal of verification in program technology development, project development, and product testing is that all work organizations must be closely integrated in the first stage of project program development. The basic tasks of unit testing of programming software development and testing application modules are mainly based on the detailed test design plan of the entire enterprise software development system, such as the detailed design plan of some data structures and analysis algorithms, etc., which are used to directly complete the software source code serial programming of a certain new type of application programming language can directly carry out the serial test of each unit after the serial programming based on each unit. At this stage of development, the technology development based on the software inspection system is related to other specific algorithm programming languages, and is implemented through a new class of technical tools including software programming, translation, application debugging, and software inspection system testing.
4.2.2. Operating Environment
The operating environment of the application software system is based on Windows. The main functional unit is implemented by the database server on the SQLServer2008 relational network and is formed by the combination of page information processing functions on a built-in Web network of SQLServer. The user-side interactive processing and transaction request and analysis processing functions use the back-end database server and can use the application program written by T.SQL to directly realize the access and analysis and calculation of all business information of the management system.
4.2.3. Program Composition
One of the main core responsibilities of the company’s customer service management information system is to quantify and manage the overall business and data of customer service. For this reason, the software first needs to be extracted from the company’s customer service management integrated business database relevant information and data, and then save the results of its processing, and then store these information and data in the enterprise’s customer service information integrated business database, so that the account manager can do statistics on the later customer service information, analysis, and other processing.
In the information system management layer based on the client and server, the corresponding application management function is to automatically access and preprocess the massive customer data stored by all customers in the customer database in real time. The basic data prediction and processing of the function call layer of each module are in the calling system, and the main function is embodied by a set of call functions applied to data processing, according to the function modules that need to be called by the upper layer and the function characteristics of the call objects of each function unit. Data-demand configuration and design have realized a set of corresponding call cases, so that other modules of the function call layer do not have to rely too much on accessing unnecessary basic analysis data, which greatly improves the efficiency of data access. It also saves a lot of time for repeated data processing for other basic analysis data processing call modules and increases the rapid accuracy of basic data processing for the entire system itself. The main core data processing architecture of this management module is a function library interface for real-time access to the database and preprocessing for warehouse management. In the comprehensive statistical data analysis that needs to be realized in SQL-based programming, we need to continue to use more complex comprehensive statistical methods for data retrieval and analysis calculations in order to effectively ensure that the statistical performance of the entire program will not be affected by the computing system. The amount of existing data in the system is largely lost, the actual program time that needs to be executed is relatively compact, and the code execution form is less.
The testing of the software is one of the key contents of the software research in this article. Different types of testing methods are used to check and verify whether the software has clear programming logic and whether the actual operation can fully meet the functional specifications expressed in the analysis, and its performance requirements and reliability. As a large-scale enterprise application software, the system’s software testing needs to be carried out in a planned way according to several stages and different levels. At the same time, it also needs to make full use of current software testing in its specific methods. Effective calculation tools in the technology and field to ensure the reliability and efficiency of its testing.
To sum up, the white box testing of the project is basically carried out on program units and operating system modules, while a dedicated testing team mainly carried out black box testing on the overall level. The testing work of the testing team focuses on verifying the various unit tests within the researcher's development team, and at the same time, it is tested and evaluated based on independent design and prepared test use cases and data, which are a core basis for testing. It is the confirmed and effective requirements analysis document. The software has been fully considered to meet various functional requirements as a whole, and the performance inspection and accuracy evaluation of the item by item have been carried out. The testing team graded and evaluated the faults found after the test according to the degree of impact on the software delivery—Table 5 In order to understand the test distance of the software, the test team listed the test content and the details of the response to the correctness of the program according to the characteristics and requirements of the system, compared with the test results and found out the cause of the failure.
5. Conclusion
With the popularization of mobile Internet and multimedia technologies, network security issues have become increasingly prominent. Image is an important signal transmission medium in the field of multimedia applications. People’s daily study, life, and work are inseparable from a large number of digital images, but these large numbers of digital images are automatically transmitted through the Internet, and they are vulnerable to malicious tampering, theft, and other threats, especially some confidential military images. Information, medical image information involving citizens’ privacy, etc., put forward stricter requirements for the protection and security of images; therefore, the most direct and effective way is to encrypt digital images. In modern cryptographic security science, two classifications are given: the basic structure of a general modern cryptographic security system and malicious attacks on new cryptographic systems. Among them, it introduces in detail all the unknown plaintext password attacks and known selected plaintext password attacks that can be selected in the plaintext password selection attack. In the assessment of the security of related technical standards for image gray-scale encryption, the technical analysis of image gray-scale histogram, differential image technology analysis, and gray-scale correlation analysis of two adjacent pixels in the image are introduced in detail. Technical criteria for evaluation in view of the company’s main business and customer management characteristics, are the information system and has two core functions: the first type of function is the core function of the system as the main service item of customer dynamic service information analysis and management, which is mainly responsible for managing the company’s beauty club. The planned and launched beauty and health care project activities and content project details, including the specific names of the project activities, content project outlines, fee payment standards and relevant preferential policies, project marketing plans, current actual project operation progress are the major data, etc. The second category is the static information of health products, because health products are not only a subsidiary unit of medical and health and beauty projects, but also one of the main ways for clubs to make profits. Therefore, it is necessary to implement a set of independent management functions for this type of information. The detailed dynamic data during the period when customers receive the service facilitates a comprehensive analysis of the satisfaction of special customers and all customers, potential demand signals and market trends, and at the same time forms a valuable operating theory for all customer managers of the company.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This paper was supported by (1) School Level: Exploration and Practice of Dual-System Talent Training Mode (no. J2017-06) and (2) Provincial Level: The Construction of Quality Evaluation System for Interdisciplinary Training of Cosmetic Technology and Engineering Talents (no. Fjjzx19-140).