Enhanced Efficiency in SMEs Attendance Monitoring: Low Cost Artificial Intelligence Facial Recognition Mobile Application

The distributed work model, which allows people to work from multiple physical places, has gained significant popularity in recent years, particularly since the outbreak of the COVID-19 epidemic. Despite the potential advantages of cost reductions associated with office space and flexible working environments, this model presents difficulties in effectively controlling and monitoring employee work activities. Applying advanced technologies in human resources can enable organizations to address these challenges effectively. However, SMEs face challenges in adopting advanced technologies such as AI and facial recognition due to financial constraints, lack of technical expertise, and concerns about cost-effectiveness. The paper focuses on developing a low-cost attendance monitoring mobile application that utilizes real-time facial recognition and location analysis for SMEs. We try to enhance the real-time registration and identification flow with the camera of employees’ mobiles by handling large data streams and integrating the system with a cloud-based solution. First, the AI facial detection module uses the camera on mobile devices as input devices to ensure the flexibility of face recognition data input. The AI Engine layer is built by applying a deep learning-based facial recognition model that was trained using a large dataset of collected employee faces. Employees’ facial identities using the Resnet34 model integrated Additive Angular Margin Loss function to identify employees’ facial identities. The facial recognition model is designed to operate in real-time, with the ability to detect and identify employees as they check in and out of the workplace. Finally, we demonstrate an experimental prototype on a mobile phone application using React Native, NodeJS, and Python. The mobile application is designed to operate on both iOS and Android platforms, ensuring widespread accessibility for SMEs. The experimental findings provide empirical evidence that our system exhibits accuracy and efficiency in the tasks of facial identification and personnel matching. This research will serve as a valuable reference for executing strategies that are appropriate for SMEs.