Pratik (2020) designed system scheduling activities in a work center. The system shall be responsible for maintaining information about employees, thus their profile. The system must include leave management from application to acceptance/rejection of leave requests, as well as all employee projects with close monitoring from creation to completion. Training will also be provided to aid in the monitoring of active and inactive employees. Making the existing system fully automatic will save a significant amount of human resources work. In the current system, all human resource work is required to record and keep the details of each and every employee in an organization. This is done so that every employee on staff can be tracked.
Simaanya (2020) conducted a study whose goal is to determine ways of improving staff management in order to provide efficiency and flexibility. The methodology used was an incremental method which helped to reduce risk when changing requirements. This research identified critical system factors that contributed most significantly to organization performance, also the research present how the system will enable HR professionals to focus on transforming information into knowledge that can be used by the organization for decision making and identified strengths and weaknesses in the organization. This research represents a first step in developing Staff management systems for HR relief. However, the system runs on centralized servers managed by the system administrator as such biometric system need to be implemented to enforce security to preserve the employees' information.
Using face recognition, Taniya et al. (2018) developed an automated HR and attendance system that can be used in real-time. In the system, recognition rates of calculations are enhanced when there are unexpected changes in a person's appearance, such as tonsured heads, wearing scarves, or whiskers. There is, however, a limitation that the system only perceives face varieties up to 30 degrees, which needs to be improved upon. Face recognition systems can be merged with step recognition systems for improved system performance.
Marcus and Rozario (2018) designed an employee database and payroll management system to make the existing manual system automatic with the help of computerized equipment and full-edged computer software, fulfilling their requirements, so that their valuable data and information can be stored for a longer period with easy access and manipulation of the same. This web application can maintain and view computerized records without getting redundant entries. The purpose of this document is to describe the functionality and specifications of the design of a web application for Managing Employees and their payroll. The expected audiences of this document are the developers and the admin of the web application. With the aid of this system, the administrator now has the data at his disposal and can simply make a solid record depending on their needs. However, there is a need to implement a facial recognition system as an additional feature to prevent system penetration from hackers even if the login credentials of the system admin are known or retrieved via hacking. Additionally, fingerprint can also be added as an authentication medium whenever data or information of a user will be modified, the essence is to ensure that the changes are made by an authorized user.
Nucleus (2018) presented research that develops a computerized human management system that may decrease repetitive work and human data input errors. This solution boosts productivity, reduces payroll errors, eliminates paper expenses, and can generate all reports on demand. Departments must manually take attendance in this system, and only this data must be input into the computerized system. However, data input errors may still occur. Staff attendance records can also be modified and faked. As a result, the suggested system must be expanded to incorporate an automatic management system to handle the latter mentioned difficulty.
The article "Adopting Biometrics for Security" by Arjun Singh (2018) identifies biometrics technology as a possible solution for controlling access to systems and networks more effectively. Despite its potential benefits, the technology has been deemed too expensive and intrusive for many organizations-and as a result adoption has been slower than expected. Although biometric authentication is more secure than traditional methods like text-based passwords, PIN numbers, and personal security questions, only 10% of respondents agree that biometrics should be the only form of authentication. One of the most common methods of biometric authentication in organizations is the use of face scanners. Approximately 14% of organizations have deployed other recognition technologies, while over half use face recognition technology.
Arulogun, Olatunbosun, Fakolujo and Olaniyi (2018) proposed a human resource management system that lays its focus on a wireless iris recognition attendance management system which was designed and implemented using Daugman's algorithm. However, building the system requires hardware for executing the iris recognition which may lead to another expense. Therefore, a final version of the system was not made.
Dey and Santhi (2017) develop a real-time facial detection-based human resource management system. Their notions were useful only when used in a variety of circumstances. These criteria included distinctive elements that could not always be controlled. For example, variation in the individual's posture, change in the brightness of the surrounding environment, and so on. Therefore, the systems were deemed wasteful if they were not used in line with needs and imperatives.
Kanchev and Kancho (2016) presented a system for developing and presenting a human resource information system for managing staff data inside a small firm or organization. It includes database and application programming features. The system has been created and is known as the Employee Management System. The developed system will be in charge of keeping records and storing data for employees inside a business, as well as creating reports as needed. Individual and specific programming tools are chosen. However, the suggested system is confined to small enterprises, and the login domain is a single mode, making the system unsecure and unreliable due to current technical improvements.
Virani & Muttu (2015) applied the Viola Jones algorithm to face detection and modified the Local Binary Pattern (LBP) algorithm to increase algorithm performance. Viola Jones algorithm was applied for face detection. The radial basis function of a neural network (NN) provided better classification accuracy than backpropagation. Despite this, more work is needed to improve facial expression recognition accuracy. According to speed tests, Histogram of Oriented Gradients (HoG) is the most efficient algorithm, followed by Haar Cascade and CNNs. In general, CNNs in Dlib perform most accurately. HoG do well for identifying faces but have a problem identifying small faces. Haar Cascade Classifiers are roughly comparable to HoG.