Towards Language-Driven Video Inpainting via Multimodal Large Language Models

TL;DR

This paper introduces a novel language-driven video inpainting task, supported by a new dataset and a diffusion-based framework that leverages multimodal large language models to perform complex, instruction-guided video editing.

Contribution

It presents the first end-to-end framework for language-guided video inpainting and introduces the ROVI dataset for training and evaluation.

Findings

The framework effectively understands and executes complex language instructions.

The ROVI dataset enables diverse inpainting scenarios.

The approach outperforms traditional mask-based methods.

Abstract

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.