Practical Considerations for Face Recognition System Implementation: Retail Business Use Case

Abstract

COVID-19 pandemic has proven experts’ prediction which stated that traditional retail might not survive in the coming age. During the pandemic, many longstanding retail brands barely survived, and some of them even went out of business. To survive and thrive, implementation of digital technologies to drive a more convenient shopping experience and enhance customer experience has proven to be crucial in gaining customer loyalty. One of the use cases that is gaining popularity nowadays is the implementation of Face Recognition system. The purpose of this study is to propose the practical solutions to three underlying issues of Face Recognition implementation: the simple but effective choice of framework, the discreet but effective way to arrange camera locations, and the light but robust choice of algorithm that could deliver good accuracy with minimum resources. This study used explorative descriptive method, combining authors’ direct experience with literature study. The result of this study is: proposed implementation framework, proposed camera arrangement, and proposed use of Neural Network algorithm with image augmentation. This study hopefully could give context to academics and fellow practitioners of the steps needed to implement Face Recognition to solve real world issues.