Object-AVEdit

An Object-level Audio-Visual Editing Model

Object-AVEdit: An Object-level Audio-Visual Editing Model

Status: Under Review for ICLR 2026
Role: Co-first Author
Affiliation: GeWu-Lab, Renmin University of China

Overview

Object-AVEdit presents a novel approach for audio-visual editing at the object level, enabling precise manipulation of audio-visual content based on object-specific controls.

Key Contributions

  • Developed an object-level audio-visual editing framework
  • Built on extensive research of diffusion-based generation and editing techniques
  • Enables precise control over audio-visual content manipulation

Authors

Youquan Fu*, Ruiyang Si*, Hongfa Wang, Dongzhan Zhou, Jiacheng Sun, Ping Luo, Di Hu, Hongyuan Zhang, Xuelong Li

* Equal contribution

Supervision

This work was conducted under the supervision of Prof. Di Hu at the GeWu Lab, Renmin University of China (February 2025 - September 2025).

This research represents my contribution to advancing multimodal AI systems, combining my interests in GenAI and unified models with practical applications in audio-visual editing.

References