OmniParser V2: Structured-Points-of-Thought for Unified Visual Text Parsing and Its Generality to Multimodal Large Language Models Paper • 2502.16161 • Published Feb 22 • 1