End-to-End Referring Video Object Segmentation with Multimodal Transformers

TL;DR

This paper introduces MTTR, a simple end-to-end multimodal Transformer model for referring video object segmentation that outperforms previous methods and simplifies the pipeline by avoiding complex components.

Contribution

The paper presents a novel Transformer-based framework for RVOS that models the task as sequence prediction, eliminating the need for complex pipelines and post-processing.

Findings

Significantly outperforms previous methods on standard benchmarks.

Achieves +5.7 and +5.0 mAP gains on A2D-Sentences and JHMDB-Sentences datasets.

Processes 76 frames per second, demonstrating efficiency.

Abstract

The referring video object segmentation task (RVOS) involves segmentation of a text-referred object instance in the frames of a given video. Due to the complex nature of this multimodal task, which combines text reasoning, video understanding, instance segmentation and tracking, existing approaches typically rely on sophisticated pipelines in order to tackle it. In this paper, we propose a simple Transformer-based approach to RVOS. Our framework, termed Multimodal Tracking Transformer (MTTR), models the RVOS task as a sequence prediction problem. Following recent advancements in computer vision and natural language processing, MTTR is based on the realization that video and text can be processed together effectively and elegantly by a single multimodal Transformer model. MTTR is end-to-end trainable, free of text-related inductive bias components and requires no additional mask-refinement post-processing steps. As such, it simplifies the RVOS pipeline considerably compared to existing methods. Evaluation on standard benchmarks reveals that MTTR significantly outperforms previous art across multiple metrics. In particular, MTTR shows impressive +5.7 and +5.0 mAP gains on the A2D-Sentences and JHMDB-Sentences datasets respectively, while processing 76 frames per second. In addition, we report strong results on the public validation set of Refer-YouTube-VOS, a more challenging RVOS dataset that has yet to receive the attention of researchers. The code to reproduce our experiments is available at https://github.com/mttr2021/MTTR