Zixiang Lu

Papers

Total Citations

H-Index

About

Zixiang Lu is a researcher specializing in human motion synthesis and multimodal machine learning, with a particular focus on generating realistic, synchronized body movements from speech. Their most-cited work, "Cospeech body motion generation using a transformer" (2024), introduces a novel transformer-based architecture that learns to produce coherent, non-verbal gestures and postures directly from audio input. This contribution addresses a critical challenge in embodied conversational agents, virtual reality, and human-computer interaction—creating natural, context-aware animations that align with spoken language. Although early in its publication cycle, the paper has already garnered attention for its elegant approach to modeling the complex temporal dependencies between speech prosody and body kinematics. Lu’s research sits at the intersection of computer vision, natural language processing, and graphics, offering practical frameworks for generating expressive avatars and digital humans. Their work promises to advance fields from assistive communication tools to immersive entertainment, and with continued development, Lu is poised to become a leading voice in the emerging domain of speech-driven motion synthesis.