End-to-End Video Instance Segmentation with Transformers

TL;DR

VisTR introduces an end-to-end Transformer-based framework for video instance segmentation, simplifying the pipeline and achieving state-of-the-art speed and competitive accuracy on the YouTube-VIS dataset.

Contribution

It presents the first simple, fast, end-to-end Transformer-based approach for VIS, framing segmentation and tracking as a sequence prediction problem.

Findings

Achieves the highest speed among VIS models.

Attains the best results among single-model methods on YouTube-VIS.

Demonstrates a simpler, faster framework with competitive accuracy.

Abstract

Video instance segmentation (VIS) is the task that requires simultaneously classifying, segmenting and tracking object instances of interest in video. Recent methods typically develop sophisticated pipelines to tackle this task. Here, we propose a new video instance segmentation framework built upon Transformers, termed VisTR, which views the VIS task as a direct end-to-end parallel sequence decoding/prediction problem. Given a video clip consisting of multiple image frames as input, VisTR outputs the sequence of masks for each instance in the video in order directly. At the core is a new, effective instance sequence matching and segmentation strategy, which supervises and segments instances at the sequence level as a whole. VisTR frames the instance segmentation and tracking in the same perspective of similarity learning, thus considerably simplifying the overall pipeline and is significantly different from existing approaches. Without bells and whistles, VisTR achieves the highest speed among all existing VIS models, and achieves the best result among methods using single model on the YouTube-VIS dataset. For the first time, we demonstrate a much simpler and faster video instance segmentation framework built upon Transformers, achieving competitive accuracy. We hope that VisTR can motivate future research for more video understanding tasks.