CityFlow-NL: Tracking and Retrieval of Vehicles at City Scale by Natural Language Descriptions
This work addresses the problem of interacting with traffic analysis systems using natural language for researchers in computer vision and urban computing, though it is incremental as it builds upon an existing benchmark.
The paper introduces CityFlow-NL, a benchmark dataset with over 5,000 natural language descriptions for vehicle targets, extending the CityFlow Benchmark to enable research on vehicle tracking and retrieval using natural language in city-scale traffic scenarios.
Natural Language (NL) descriptions can be one of the most convenient or the only way to interact with systems built to understand and detect city scale traffic patterns and vehicle-related events. In this paper, we extend the widely adopted CityFlow Benchmark with NL descriptions for vehicle targets and introduce the CityFlow-NL Benchmark. The CityFlow-NL contains more than 5,000 unique and precise NL descriptions of vehicle targets, making it the first multi-target multi-camera tracking with NL descriptions dataset to our knowledge. Moreover, the dataset facilitates research at the intersection of multi-object tracking, retrieval by NL descriptions, and temporal localization of events. In this paper, we focus on two foundational tasks: the Vehicle Retrieval by NL task and the Vehicle Tracking by NL task, which take advantage of the proposed CityFlow-NL benchmark and provide a strong basis for future research on the multi-target multi-camera tracking by NL description task.