Towards Language-Driven Video Inpainting via Multimodal Large Language Models
This addresses the tedious and labor-intensive process of manual mask labeling in video inpainting for researchers and practitioners in computer vision.
The paper tackles the problem of video inpainting by introducing a language-driven approach that uses natural language instructions instead of manual binary masks, resulting in the creation of the ROVI dataset with 5,650 videos and 9,091 inpainting results and a novel diffusion-based framework as the first end-to-end baseline.
We introduce a new task -- language-driven video inpainting, which uses natural language instructions to guide the inpainting process. This approach overcomes the limitations of traditional video inpainting methods that depend on manually labeled binary masks, a process often tedious and labor-intensive. We present the Remove Objects from Videos by Instructions (ROVI) dataset, containing 5,650 videos and 9,091 inpainting results, to support training and evaluation for this task. We also propose a novel diffusion-based language-driven video inpainting framework, the first end-to-end baseline for this task, integrating Multimodal Large Language Models to understand and execute complex language-based inpainting requests effectively. Our comprehensive results showcase the dataset's versatility and the model's effectiveness in various language-instructed inpainting scenarios. We will make datasets, code, and models publicly available.