Two papers were accepted at this year’s RoboLetics: Workshop on Robot Learning in Athletics@ CoRL 2023. The papers are titled Optimizing Long-Term Player Tracking and Identification in NAO Robot Soccer by fusing Game-state and External Video and Multi-Domain Referee Dataset: Enabling Recognition of Referee Signals on Robotic Platforms. The first paper uses the game state and external video for tracking robots across multiple frames, while the second one introduces a dataset and baseline model for this year’s RoboCup in-game challenge of tracking visual referee signals.
Figure 2: Chrome key data collection and augmentation from Multi-Domain Referee Dataset.