Generation of virtual digital human for customer service industry

With the widespread development of digital technology, individuals' daily activities are inseparable from interaction with electronic devices. Researchers have become interested in developing novel methods, to enable users to experience social and emotional satisfaction that traditional face-to-face interaction provides. In this study, we propose a novel deep learning-based pipeline to generate virtual digital humans for customer service industry. Specifically, we propose a method to construct a database with template service actions. Furthermore, we propose a two-stage method for generating 2D virtual human videos with gestures and emotional lip-sync expressions. We have conducted qualitative and quantitative experiments on the proposed 2D virtual human video generation method. The results demonstrate that the method effectively generates high-quality virtual digital humans for the customer service industry.& COPY; 2023 Elsevier Ltd. All rights reserved.