SIS-Challenge: Event-based Spatio-temporal Instance Segmentation Challenge at the CVPR 2025 Event-based Vision Workshop

TL;DR

The SIS-Challenge at CVPR 2025 focused on advancing spatio-temporal instance segmentation by evaluating methods that predict pixel-level masks from event-based and grayscale camera data, fostering progress in event-based vision.

Contribution

This paper introduces the SIS challenge, providing datasets, evaluation protocols, and insights into top-performing methods for event-based spatio-temporal segmentation.

Findings

Top methods achieved high segmentation accuracy.

Event-based data can be effectively used for real-time segmentation.

The challenge spurred diverse innovative approaches.

Abstract

We present an overview of the Spatio-temporal Instance Segmentation (SIS) challenge held in conjunction with the CVPR 2025 Event-based Vision Workshop. The task is to predict accurate pixel-level segmentation masks of defined object classes from spatio-temporally aligned event camera and grayscale camera data. We provide an overview of the task, dataset, challenge details and results. Furthermore, we describe the methods used by the top-5 ranking teams in the challenge. More resources and code of the participants' methods are available here: https://github.com/tub-rip/MouseSIS/blob/main/docs/challenge_results.md