Dataset Management Platform

Develop a platform for managing spatiotemporal datasets of road scenes, with functions including data cleaning, annotation management, and dataset organization and storage. The platform needs to be able to process data in image and video formats, and simultaneously conduct comprehensive management of road elements, road participants, events, etc. contained therein.

Spatio-temporal large model

Design and train a spatio-temporal large model for road scenarios. This model can generate synthetic data, capture complex dynamic information in various aspects such as road elements, traffic participants, and weather changes, and produce stable and accurate images and videos.

Synthetic Data Platform

Build a synthetic data platform where users can flexibly specify scenarios, road elements, weather conditions, road participants, events, etc., to generate images and videos. At the same time, it provides detailed annotations for the generated content to support the training of intelligent driving systems.