Video Instance Segmentation using Inter-Frame Communication Transformers

TL;DR

This paper introduces Inter-frame Communication Transformers (IFC), a new efficient transformer-based approach for video instance segmentation that achieves state-of-the-art accuracy with high speed and low computational overhead.

Contribution

The paper presents IFC, a novel transformer architecture with memory tokens that efficiently encodes inter-frame context for improved video instance segmentation.

Findings

Achieved state-of-the-art AP 44.6 on YouTube-VIS 2019 val set.

Processed videos at 89.4 FPS, enabling real-time inference.

Reduced computational overhead compared to previous per-clip models.

Abstract

We propose a novel end-to-end solution for video instance segmentation (VIS) based on transformers. Recently, the per-clip pipeline shows superior performance over per-frame methods leveraging richer information from multiple frames. However, previous per-clip models require heavy computation and memory usage to achieve frame-to-frame communications, limiting practicality. In this work, we propose Inter-frame Communication Transformers (IFC), which significantly reduces the overhead for information-passing between frames by efficiently encoding the context within the input clip. Specifically, we propose to utilize concise memory tokens as a mean of conveying information as well as summarizing each frame scene. The features of each frame are enriched and correlated with other frames through exchange of information between the precisely encoded memory tokens. We validate our method on the latest benchmark sets and achieved the state-of-the-art performance (AP 44.6 on YouTube-VIS 2019 val set using the offline inference) while having a considerably fast runtime (89.4 FPS). Our method can also be applied to near-online inference for processing a video in real-time with only a small delay. The code will be made available.