Video Editing Chatbot: Language-Driven Video Compositing System

摘要

In this work, we present a video editing chatbot (VEC) that performs intelligent multimedia editing through natural language dialogue. VEC comprises three modules: instruction analysis, multimedia resources retrieval, and multimedia resources editing. It analyzes user instructions to retrieve relevant multimedia resources from the multimedia database (MMDB), and then applies appropriate editing methods from the multimedia toolbase (MMTB) automatically. To enhance user experience and simplify operation, VEC uses a multi-turn dialogue mechanism to handle complex editing tasks.