TUMTraf EMOT: A New Dataset and Baseline for Event-Based Multi-Object Tracking in Traffic

The paper introduces a new dataset and baseline for multi-object tracking using event-based vision in traffic scenarios, which is a promising research area. Event-based vision offers potential advantages in challenging lighting and speed conditions compared to traditional methods.

• •Presents a new dataset called TUMTraf EMOT.
• •Develops a baseline for event-based multi-object tracking.
• •Addresses the problem of object tracking in traffic using event cameras.