Tang focuses on sign language video translation and generation research, involving computer science, artificial intelligence, machine learning, linguistics, and other fields. The main research directions include sign language video data processing and recognition, conversion between sign language videos and spoken languages, natural and smooth sign language video generation, etc.

Sign language video translation and generation is a research topic that aims to use artificial intelligence technology to provide a better communication experience for the deaf community. This field focuses on how to convert spoken language into sign language videos, and how to convert sign language videos into spoken language. The goal of this task of research is to use computer algorithms and machine learning techniques to automatically recognize gestures and facial expressions in sign language videos and then convert them into spoken language or other forms of text output. Besides, the field is also working to generate natural, fluent sign language videos so that spoken language can be more easily understood by the deaf community.

In this field, researchers can explore how to design effective algorithms to process sign language video data, how to train neural networks to automatically recognize sign language, and how to leverage natural language processing techniques to generate natural and fluent sign language videos. At the same time, they can also study the language and culture of the deaf community to better understand their needs and provide them with better services.